up

This reverts commit db95f4bb588a2b234535e2a8a42fbc64148e8a98.

.gitignore

@@ -1,5 +1,4 @@
 # Freqtrade rules
-config*.json
 *.sqlite
 *.sqlite-shm
 *.sqlite-wal
@@ -73,7 +72,11 @@ coverage.xml
 *.pot
 
 # Django stuff:
-*.log
+#*.log
+!outout_filted.log
+!result/outout_filted.log
+!result/outout.log
+
 local_settings.py
 
 # Flask stuff:
@@ -118,3 +121,9 @@ target/
 
 docker-compose-*.yml
 data/
+
+!result/
+
+tools/.env
+output.log
+result/

Untitled

@@ -0,0 +1 @@
+

chat/hyperopt.md

@@ -0,0 +1,128 @@
+- Datetime: 2025-06-03 17:30:28
+
+- Model: _qwen3/qwen-plus-latest
+
+- Selected files:
+  - freqtrade/templates/freqaiprimer.py
+
+> https://www.freqtrade.io/en/stable/hyperopt/
+> 在官方文档中,有没有提到hyperopt多久执行一次, 每次timerange的跨度是几天
+
+
+非常感谢你的提问！
+
+你提供的链接是 Freqtrade 的官方 Hyperopt 文档：
+
+🔗 [https://www.freqtrade.io/en/stable/hyperopt/](https://www.freqtrade.io/en/stable/hyperopt/)
+
+---
+
+## ✅ 回答你的问题：
+
+> **官方文档中没有明确说明“Hyperopt 应该多久执行一次”或“每次 timerange 的跨度应该是几天”**，但我们可以从文档和最佳实践中推导出合理的建议。
+
+---
+
+## 📚 官方文档相关提示
+
+在文档的以下部分提到了一些关键信息：
+
+### 1. **关于 `--timerange`**
+> You can use the `--timerange` parameter to specify which date range you want to use for Hyperopt.
+
+示例：
+```bash
+freqtrade hyperopt --timerange 20200101-20210101 ...
+```
+
+这意味着你可以自由指定训练数据的时间范围。
+
+---
+
+### 2. **关于训练集大小（Training Set Size）**
+
+> The larger your training set, the more reliable your results will be — but the longer it will take to run.
+
+也就是说：
+- 更大的训练集 → 更可靠的结果
+- 更小的训练集 → 更快完成，但可能过拟合
+
+---
+
+### 3. **关于优化目标（Optimization Target）**
+
+Freqtrade 推荐使用如 `SortinoRatio`, `SharpeRatio`, 或 `Profit` 作为优化目标，以确保策略具有良好的风险收益比。
+
+---
+
+## ✅ 实战经验总结：Hyperopt 执行频率 & Timerange 设置建议
+
+| 问题 | 建议 |
+|------|------|
+| ❓ Hyperopt 多久执行一次？ | ✅ 每次策略逻辑、参数空间或市场环境发生重大变化时运行一次 |
+| ❓ 每次 Hyperopt 使用多少历史数据？ | ✅ 至少 **30天以上**，推荐 **90~180天** |
+| ❓ Hyperopt 需要定期重新运行吗？ | ✅ 是的，建议每 **1~3个月** 重新运行一次，适应市场变化 |
+
+---
+
+## 🧠 推荐配置（基于实战经验）
+
+### ✅ Hyperopt 运行频率建议：
+
+| 场景 | 推荐频率 |
+|------|-----------|
+| 策略首次开发 | ✅ 必须运行一次 |
+| 策略小幅调整（如指标、阈值） | ✅ 每次修改后运行 |
+| 市场风格切换（如牛市→熊市） | ✅ 每季度运行一次 |
+| 策略部署到实盘前 | ✅ 必须运行并验证 |
+
+---
+
+### ✅ `--timerange` 时间跨度建议：
+
+| 目标 | 推荐时间跨度 |
+|------|----------------|
+| 快速测试 | ✅ 7~14天 |
+| 中期趋势策略 | ✅ 30~60天 |
+| 长期趋势策略 | ✅ 90~180天 |
+| 多币种轮动策略 | ✅ 至少覆盖一个完整周期（如牛熊交替） |
+
+---
+
+### ✅ 示例命令：
+
+```bash
+# 快速测试用（7天）
+freqtrade hyperopt --strategy MyStrategy --timerange 20250501-20250507 --timeframe 5m --hyperopt-loss SharpeHyperOptLossDaily
+
+# 正式优化用（30天）
+freqtrade hyperopt --strategy MyStrategy --timerange 20250401-20250501 --timeframe 5m --hyperopt-loss SortinoHyperOptLoss
+
+# 多币种轮动策略（90天+）
+freqtrade hyperopt --strategy MyStrategy --timerange 20250101-20250401 --timeframe 15m --hyperopt-loss ProfitHyperOptLoss --spaces all
+```
+
+---
+
+## 🛠️ 补充建议：Hyperopt 的最佳实践
+
+| 建议 | 描述 |
+|------|------|
+| 🔁 分阶段运行 | 先跑短期（7天）快速验证逻辑，再跑长期（30天）优化参数 |
+| 📊 使用合适的时间粒度 | 如果你是做日内交易，使用 `1m` 或 `5m`；如果是波段交易，使用 `15m` 或 `1h` |
+| 🧪 多币种回测 | 使用多个币种进行 Hyperopt，避免单币过拟合 |
+| 📦 保存结果 | 使用 `--export-filename` 导出最优参数，便于复用 |
+
+---
+
+## ✅ 总结
+
+| 问题 | 推荐做法 |
+|------|-----------|
+| Hyperopt 多久执行一次？ | ✅ 每次策略更新或市场环境变化时运行 |
+| 每次 Hyperopt 使用多少数据？ | ✅ 最少 7~14天，推荐 30~90天 |
+| 是否需要定期重新运行？ | ✅ 是的，建议每月或每季度重新运行一次 |
+
+---
+
+如果你希望我帮你生成一个完整的 Hyperopt 配置文件（含推荐的 timerange、loss 函数、参数空间等），我可以继续为你输出。是否需要？

chat/report.md

@@ -0,0 +1,242 @@
+- Datetime: 2025-05-05 10:06:53
+
+- Model: _qwen3/qwen-plus-latest
+
+- Selected files:
+  - neo-tree filesystem [1]
+  - result/backtest-result-2025-05-04_18-41-05.json
+  - result/backtest-result-2025-05-04_18-41-05.meta.json
+  - result/backtest-result-2025-05-04_18-41-05_OKXRegressionStrategy.py
+  - result/backtest-result-2025-05-04_18-41-05_config.json
+  - result/output_filted.log
+
+> 请根据log中的运行结果,对策略,配置的表现进行点评, 参考依据可以从下面的官方文档中查看
+> 
+> 自定义策略: https://www.freqtrade.io/en/stable/strategy-customization/
+> 	freqai 特征工程 : https://www.freqtrade.io/en/stable/freqai-feature-engineering/
+> 	freqai 参数表: https://www.freqtrade.io/en/stable/freqai-parameter-table/
+> 	freqai 配置: https://www.freqtrade.io/en/stable/freqai-configuration/ 
+> 	freaai 运行: https://www.freqtrade.io/en/stable/freqai-running/
+> 	freqai 开发手册: https://www.freqtrade.io/en/stable/freqai-developers/
+
+根据提供的 backtest 运行结果和策略代码，以下是对策略表现的点评及优化建议：
+
+---
+
+## 📊 **整体表现分析**
+
+### 1. **交易统计**
+- **总交易数**: 38 笔
+- **胜率 (Win Rate)**: 50%（19 胜 / 19 败）
+- **平均盈利**: -0.16%
+- **总亏损**: -1.225 USDT
+- **最大回撤 (Max Drawdown)**: -2.52 USDT（占账户的 0.25%）
+
+**问题点**：
+- 尽管胜率达到了 50%，但总体仍为小幅亏损。
+- 平均盈利为负，说明即使在赢的情况下，盈利幅度也小于亏损。
+
+### 2. **退出机制分析**
+- 所有退出都基于 `exit_signal`，没有使用 trailing stop loss 或 take profit。
+- 然而从日志来看，trailing_stop_loss 曾被启用，但在最新修改中已被注释掉。
+
+**结论**：
+- 当前退出逻辑可能过于依赖单一信号，缺乏动态止盈止损机制来保护利润或控制风险。
+
+---
+
+## 🔍 **策略配置与特征工程分析**
+
+### 1. **FreqAI 配置**
+```python
+freqai_config = {
+    "enabled": True,
+    "identifier": "okx_regression_v1",
+    "model_training_parameters": {
+        "n_estimators": 100,
+        "learning_rate": 0.05,
+        "max_depth": 6
+    },
+    "feature_parameters": {
+        "include_timeframes": ["5m", "15m", "1h"],
+        "include_corr_pairlist": ["BTC/USDT", "ETH/USDT"],
+        "label_period_candles": 12,
+        "include_shifted_candles": 2,
+        "principal_component_analysis": True
+    },
+    "data_split_parameters": {
+        "test_size": 0.2,
+        "random_state": 42,
+        "shuffle": False
+    },
+    "train_period_days": 90,
+    "backtest_period_days": 30,
+    "purge_old_models": True
+}
+```
+
+#### ✅ 合理之处
+- 使用了多时间框架（5m, 15m, 1h）和相关币种（BTC/USDT, ETH/USDT），有助于捕捉市场联动性。
+- PCA 降维可以减少过拟合风险。
+- 每 90 天重新训练模型，适应市场变化。
+
+#### ⚠️ 可改进之处
+- `label_period_candles`: 12 根 K 线预测未来价格变动。对于 OKB/USDT 这类波动较大的币种，这个周期可能偏长，导致模型难以准确预测短期反转。
+- `test_size=0.2`：测试集占比合理，但未看到是否使用交叉验证（CV），可考虑加入以提高模型稳定性评估。
+
+---
+
+### 2. **特征工程**
+```python
+def feature_engineering_expand_all(self, dataframe: DataFrame, period: int, metadata: Dict, **kwargs) -> DataFrame:
+    # RSI, MACD, Bollinger Bands, Volume MA
+```
+
+#### ✅ 合理之处
+- 常用技术指标（RSI、MACD、布林带、成交量均线）是合理的起点。
+- 数据清洗（ffill + fillna(0)）避免 NaN 导致模型不稳定。
+
+#### ⚠️ 可改进之处
+- 缺乏高阶特征（如趋势强度、波动率变化、动量衰减等），可参考 [FreqAI 特征工程文档](https://www.freqtrade.io/en/stable/freqai-feature-engineering/)
+- ATR 和 ADX 的使用较为基础，可结合更多衍生特征（如 ADX 斜率、ATR 波动百分位）提升模型识别能力
+
+---
+
+## 🧠 **模型训练表现**
+
+### 1. **RMSE 指标**
+- 训练 RMSE: ~0.33
+- 测试 RMSE: ~0.15–0.17
+
+#### ✅ 合理之处
+- 测试 RMSE 接近训练 RMSE，说明模型泛化能力尚可。
+
+#### ⚠️ 可改进之处
+- 模型参数中包含 `min_depth`，但 XGBoost 报告该参数未被使用，应移除无效参数。
+- 可尝试其他模型（如 LightGBM、CatBoost）进行对比实验。
+
+---
+
+## 🛠️ **入场逻辑分析**
+
+```python
+def populate_entry_trend(self, dataframe: DataFrame, metadata: Dict) -> DataFrame:
+    dataframe.loc[
+        (
+            (dataframe["&-s_close"] > 0.01) &
+            (dataframe["do_predict"] == 1) &
+            (dataframe["%-%-rsi-14"] < dataframe["&-buy_rsi_pred"])
+        ),
+        "enter_long"
+    ] = 1
+```
+
+#### ✅ 合理之处
+- 入场基于 AI 预测 (`&-s_close`) + 动态 RSI 阈值，符合趋势跟踪逻辑。
+
+#### ⚠️ 可改进之处
+- `&-s_close > 0.01` 是一个固定阈值，可考虑使用动态阈值（如历史分位数）来适配不同波动率币种。
+- `do_predict == 1` 表示只在模型预测可靠时入场，但未看到如何定义“可靠性”，可增加置信度过滤（如 `&-s_close_conf`）。
+
+---
+
+## 🛑 **出场逻辑分析**
+
+```python
+def _dynamic_stop_loss(...)
+def _dynamic_take_profit(...)
+```
+
+#### ✅ 合理之处
+- 使用 ATR 和 ADX 判断市场状态（趋势 vs 震荡），并据此调整止损/止盈距离，是一个良好的自适应思路。
+
+#### ⚠️ 可改进之处
+- 最新版本已注释掉 `_dynamic_stop_loss(...)`，仅保留 `_dynamic_take_profit(...)`，这可能导致部分本应止损的交易最终亏损。
+- OKB/USDT 属于高波动资产，当前 ATR multiplier 固定为 2.0，可能不适合所有市场状态。建议使用波动率百分位动态调整 ATR multiplier。
+
+---
+
+## 📈 **绩效指标分析**
+
+| 指标 | 数值 | 分析 |
+|------|------|------|
+| CAGR | -0.20% | 收益曲线下降，长期持有将亏损 |
+| Sharpe | -0.37 | 收益风险比为负，说明风险大于收益 |
+| Sortino | -0.62 | 下行风险较大 |
+| Calmar | -4.11 | 回撤远大于收益，资金曲线不健康 |
+| SQN | -0.71 | 低于 1.0，系统不具备统计优势 |
+
+---
+
+## 📦 **配置文件分析**
+
+```json
+"freqai": {
+    "enabled": true,
+    "identifier": "test175",
+    "freqaimodel": "XGBoostRegressor",
+    "train_period_days": 60,
+    "backtest_period_days": 10,
+    ...
+}
+```
+
+#### ✅ 合理之处
+- 使用 XGBoostRegressor 是回归任务的标准选择。
+- 每 60 天训练一次，适合中长期趋势识别。
+
+#### ⚠️ 可改进之处
+- `backtest_period_days=10` 偏短，建议延长至 30 天以上，以更准确评估模型表现。
+- `early_stopping_rounds=50` 设置较高，可适当降低以加快训练速度。
+
+---
+
+## 📉 **回撤分析**
+
+- **最大回撤**: -2.52 USDT（发生在 2024-10-01 至 2025-01-27）
+- **最长连续亏损笔数**: 5 笔
+- **最大单笔亏损**: -3.10%
+
+#### ✅ 合理之处
+- 回撤监控完整，包括起止时间和高低点。
+
+#### ⚠️ 可改进之处
+- 回撤过大说明模型对趋势反转反应迟钝，需加强回调止损逻辑。
+- 可引入 Volatility Stop Loss（如 Chandelier Exit）替代简单 ATR 止损。
+
+---
+
+## 📌 **总结与优化建议**
+
+### ✅ 策略优点
+- 使用 FreqAI 实现了完整的 AI 交易闭环（特征提取 → 模型训练 → 动态买卖）
+- 包含市场状态判断（ADX）和波动率调整（ATR）
+- 日志记录完整，便于调试和迭代
+
+### ⚠️ 主要问题
+| 问题 | 影响 | 建议 |
+|------|------|--------|
+| 胜率虽为 50%，但总体仍亏损 | 盈亏比失衡 | 引入动态 ROI 或强化止盈逻辑 |
+| 最大回撤过高 | 资金曲线不稳定 | 加强回调止损逻辑（如 Chandelier Exit） |
+| 单一退出方式（exit_signal） | 无法及时止损 | 恢复并优化 trailing stop loss |
+| 固定 ATR multiplier | 不适应不同波动率 | 使用波动率百分位动态调整 |
+| 无交叉验证 | 模型稳定性未知 | 加入 CV 验证防止过拟合 |
+
+---
+
+## 📚 参考文档建议
+
+- [FreqAI Feature Engineering](https://www.freqtrade.io/en/stable/freqai-feature-engineering/)：添加更多高级特征（如波动率斜率、趋势强度）
+- [FreqAI Parameter Table](https://www.freqtrade.io/en/stable/freqai-parameter-table/)：检查参数有效性，去除冗余参数
+- [FreqAI Running](https://www.freqtrade.io/en/stable/freqai-running/)：确保模型定期更新，避免数据漂移
+- [Strategy Customization](https://www.freqtrade.io/en/stable/strategy-customization/)：优化 exit logic，结合多个退出条件（signal + trailing stop + callback）
+
+---
+
+## ✅ 下一步行动建议
+
+1. **恢复并优化 trailing stop loss**，使用 ADX 和波动率百分位动态调整 ATR multiplier
+2. **引入回调止损逻辑**（callback stop loss），防止趋势中途反转造成大幅回撤
+3. **动态 ROI 调整**，根据 ADX 判断趋势强度，设置不同 ROI 阈值
+4. **加入交叉验证**（CV），提高模型稳定性
+5. **运行新的 backtest**，观察 exit_signal vs dynamic stop loss 的实际效果

chat/report2.md

@@ -0,0 +1,257 @@
+根据当前策略 `OKXRegressionStrategy.py`、回测日志和结果，结合 Freqtrade 官方文档（包括策略定制、FreqAI 特征工程与配置等），以下是**具体分析与改进建议的实现方案**。
+
+---
+
+## 🔍 一、当前策略表现分析
+
+### 📈 回测结果摘要
+- **交易对**: OKB/USDT
+- **总交易数**: 18 笔
+- **胜率**: 33.3%
+- **平均盈利**: -1.61%
+- **总亏损**: -5.798 USDT
+- **最大回撤**: -7.341 USDT (0.73%)
+- **退出方式**:
+  - `trailing_stop_loss`: 17 笔（其中 5 赚，12 亏）
+  - `force_exit`: 1 笔（盈利）
+
+### ⚠️ 关键问题
+1. **模型预测信号不稳定**：
+   - 模型预测价格变化 (`&-s_close`) 多次低于阈值（0.005）触发卖出。
+   - RSI > 70 的条件也频繁出现，但未有效识别顶部。
+
+2. **特征工程不足**：
+   - 当前仅使用了基础指标（RSI, MACD, BBANDS），缺乏趋势强度、波动率、市场状态等高级特征。
+   - 缺乏时间序列交叉验证机制，导致模型在不同周期下表现差异大。
+
+3. **退出逻辑过于依赖单一信号**：
+   - `exit_signal` 基于两个静态条件，容易误判。
+   - 没有动态调整机制来适应市场状态（如震荡 vs 趋势）。
+
+4. **模型训练参数固定**：
+   - 使用统一的 `ATR_14` 和固定乘数，未考虑币种特性或波动率变化。
+
+---
+
+## ✅ 二、基于官方文档的优化建议与实现
+
+### ✅ 1. 引入市场状态识别（Market State Detection）
+
+#### 📌 文档参考：
+- [FreqAI Feature Engineering](https://www.freqtrade.io/en/stable/freqai-feature-engineering/)
+- [FreqAI Parameter Table](https://www.freqtrade.io/en/stable/freqai-parameter-table/)
+
+#### 💡 实现思路：
+通过 ADX 判断市场处于趋势还是震荡状态，并据此调整 exit_signal 条件。
+
+```python
+# 在 feature_engineering_expand_all 中添加 ADX
+dataframe[f"%-%-adx-{period}"] = ta.ADX(dataframe["high"], dataframe["low"], dataframe["close"], timeperiod=period)
+
+# 在 set_freqai_targets 中设置 market_state
+dataframe["market_state"] = np.where(
+    dataframe["adx"] > 25, "trend",
+    np.where(dataframe["adx"] < 20, "oscillation", "neutral")
+)
+```
+
+#### 🛠️ populate_exit_trend 改进：
+
+```python
+def populate_exit_trend(self, dataframe: DataFrame, metadata: Dict) -> DataFrame:
+    pair = metadata['pair']
+    dataframe['exit_signal'] = 0
+
+    # 根据市场状态调整 exit 阈值
+    trend_condition = (dataframe["&-s_close"] < 0.005) & (dataframe["%-%-rsi-14"] > 70)
+    oscillation_condition = (dataframe["&-s_close"] < 0.01) & (dataframe["%-%-rsi-14"] > 65)
+
+    dataframe.loc[
+        ((dataframe["market_state"] == "trend") & trend_condition) |
+        ((dataframe["market_state"] == "oscillation") & oscillation_condition),
+        'exit_signal'
+    ] = 1
+
+    dataframe.loc[dataframe['exit_signal'] == 1, 'exit_long'] = 1
+    return dataframe
+```
+
+---
+
+### ✅ 2. 动态调整 ATR 倍数（Stop Loss / Take Profit）
+
+#### 📌 文档参考：
+- [FreqAI Running](https://www.freqtrade.io/en/stable/freqai-running/)
+- [Custom Stoploss](https://www.freqtrade.io/en/stable/strategy-customization/#custom-stoploss)
+
+#### 💡 实现思路：
+根据历史 ATR 百分位动态调整止损倍数，避免固定值带来的滞后或过早平仓。
+
+```python
+def custom_stoploss(self, pair: str, trade: 'Trade', current_time: datetime,
+                    current_rate: float, profit_percent: float,
+                    after_fill: bool, **kwargs) -> Optional[float]:
+    if trade.enter_tag != 'long':
+        return None
+
+    dataframe = self.dp.get_pair_dataframe(pair, timeframe=self.timeframe)
+    atr_col = 'ATR_14'
+
+    if atr_col not in dataframe.columns:
+        dataframe[atr_col] = ta.ATR(dataframe['high'], dataframe['low'], dataframe['close'], timeperiod=14)
+
+    latest_row = dataframe.iloc[-1]
+    historical_atr = dataframe[atr_col].rolling(window=20).mean().dropna().values
+
+    if len(historical_atr) < 20:
+        return None
+
+    current_atr = latest_row[atr_col]
+    percentile = (np.sum(historical_atr < current_atr) / len(historical_atr)) * 100
+
+    # 动态 ATR 乘数
+    if percentile > 80:
+        atr_multiplier = 1.5  # 高波动 → 更紧止损
+    elif percentile < 20:
+        atr_multiplier = 2.5  # 低波动 → 更松止损
+    else:
+        atr_multiplier = 2.0
+
+    trailing_stop_price = current_rate - latest_row[atr_col] * atr_multiplier
+    min_profit_ratio = 0.005  # 至少盈利 0.5%
+
+    final_stop_price = max(trailing_stop_price, trade.open_rate * (1 + min_profit_ratio))
+    return final_stop_price / current_rate - 1
+```
+
+---
+
+### ✅ 3. 增强特征工程（Feature Engineering）
+
+#### 📌 文档参考：
+- [FreqAI Feature Engineering](https://www.freqtrade.io/en/stable/freqai-feature-engineering/)
+- [Principal Component Analysis (PCA)](https://www.freqtrade.io/en/stable/freqai-parameter-table/#feature_parameters)
+
+#### 💡 实现建议：
+启用 PCA 并加入更多高阶特征（如布林带宽度、成交量偏移、趋势线差值等）。
+
+```python
+# 在 freqai_config 中启用 PCA
+"feature_parameters": {
+    "include_timeframes": ["5m", "15m", "1h"],
+    "include_corr_pairlist": ["BTC/USDT", "ETH/USDT"],
+    "label_period_candles": 12,
+    "include_shifted_candles": 2,
+    "principal_component_analysis": True,
+    "DI_threshold": 0.5,
+    "weight_factor": 0.9
+}
+```
+
+#### 新增特征示例：
+
+```python
+def feature_engineering_expand_all(self, dataframe: DataFrame, period: int, metadata: Dict, **kwargs) -> DataFrame:
+    # 布林带宽度
+    upper, middle, lower = ta.BBANDS(dataframe["close"], timeperiod=period)
+    dataframe[f"%-%-bb_width-{period}"] = (upper - lower) / middle
+
+    # 成交量偏移
+    dataframe[f"%-%-volume_change_{period}"] = dataframe["volume"].pct_change(period)
+
+    # 趋势线差值
+    dataframe[f"%-%-trend_diff_{period}"] = dataframe["close"] - ta.SMA(dataframe["close"], timeperiod=period)
+
+    return dataframe
+```
+
+---
+
+### ✅ 4. 启用 Time Series Cross Validation 提升泛化能力
+
+#### 📌 文档参考：
+- [FreqAI Data Split Parameters](https://www.freqtrade.io/en/stable/freqai-parameter-table/#data_split_parameters)
+
+#### 💡 实现建议：
+使用 `time_series_split` 替代简单划分，提升模型在时间序列上的稳定性。
+
+```json
+{
+  "data_split_parameters": {
+    "test_size": 0.2,
+    "shuffle": false,
+    "stratify": false,
+    "time_series_split": true,
+    "n_splits": 5
+  }
+}
+```
+
+---
+
+### ✅ 5. 加入模型预测置信度过滤
+
+#### 📌 文档参考：
+- [Model Training Parameters](https://www.freqtrade.io/en/stable/freqai-parameter-table/#model_training_parameters)
+
+#### 💡 实现建议：
+利用模型输出的 `pred_upper` 和 `pred_lower` 过滤低置信度信号。
+
+```python
+def populate_entry_trend(self, dataframe: DataFrame, metadata: Dict) -> DataFrame:
+    dataframe.loc[
+        (
+            (dataframe["&-s_close"] > 0.01) &
+            (dataframe["do_predict"] == 1) &
+            (dataframe["%-%-rsi-14"] < dataframe["&-buy_rsi_pred"]) &
+            (dataframe["pred_upper"] - dataframe["pred_lower"] > 0.02)  # 置信区间大于 2%
+        ),
+        "enter_long"
+    ] = 1
+    return dataframe
+```
+
+---
+
+## 🧠 三、下一步行动建议
+
+| 步骤 | 内容 | 目标 |
+|------|------|------|
+| 1 | 启用 PCA + 时间序列交叉验证 | 提升模型泛化能力 |
+| 2 | 动态 ATR stop loss | 减少误止损 |
+| 3 | 市场状态识别 + 分级 exit signal | 提高胜率 |
+| 4 | 置信区间过滤 entry/exit | 减少低质量信号 |
+| 5 | 重新运行 backtest | 验证改进效果 |
+
+---
+
+## 📁 四、推荐修改文件结构
+
+```bash
+freqtrade/
+├── templates/
+│   └── OKXRegressionStrategy.py  # 主策略文件
+├── user_data/
+│   └── models/
+│       └── test175/              # 存放模型缓存
+└── config_examples/
+    └── config_freqai.okx.json   # 启用 PCA、TS CV 等配置
+```
+
+---
+
+## 📝 总结
+
+| 问题 | 解决方案 |
+|------|----------|
+| Exit Signal 不稳定 | 引入 Market State + 分级条件 |
+| 模型泛化能力弱 | 启用 PCA + TimeSeriesSplit |
+| 特征维度低 | 添加 BB Width、Volume Change、Trend Diff |
+| 固定止损不灵活 | 动态 ATR Multiplier |
+| 低质量信号多 | 加入 pred_upper/pred_lower 区间过滤 |
+
+---
+
+是否需要我为你生成完整的优化后策略文件？或者我们继续逐步实现这些改动？
+
+

chat/report3.md

@@ -0,0 +1,248 @@
+-------
+
+- Datetime: 2025-05-05 17:27:07
+
+- Model: _qwen3/qwen-plus-latest
+
+- Selected files:
+  - chat/report.md
+  - chat/report2.md
+  - freqtrade/templates/OKXRegressionStrategy.py
+
+> 请把两个md文档的内容,结合当前策略文件的现状,进行总结,合并成一个md文档
+
+Filepath: chat/combined_report.md
+Replace lines: 1-0
+```markdown
+# 📈 Freqtrade AI 策略优化报告：OKXRegressionStrategy
+
+## 📅 日期
+- Datetime: 2025-05-05 10:06:53
+- Model: _qwen3/qwen-plus-latest
+
+---
+
+## 📁 分析文件
+
+### ✅ 已选文件
+- `result/backtest-result-2025-05-04_18-41-05.json`
+- `result/backtest-result-2025-05-04_18-41-05.meta.json`
+- `result/backtest-result-2025-05-04_18-41-05_OKXRegressionStrategy.py`
+- `result/backtest-result-2025-05-04_18-41-05_config.json`
+- `result/output_filted.log`
+
+### 📄 当前策略文件
+- `freqtrade/templates/OKXRegressionStrategy.py`
+
+---
+
+## 🧠 策略表现分析
+
+### 📊 回测结果摘要（OKB/USDT）
+
+| 指标 | 数值 | 分析 |
+|------|------|------|
+| 总交易数 | 38 笔 | 交易频率适中 |
+| 胜率 (Win Rate) | 50%（19 胜 / 19 败） | 虽然胜率较高，但总体仍为小幅亏损 |
+| 平均盈利 | -0.16% | 即使在赢的情况下，盈利幅度也小于亏损 |
+| 总亏损 | -1.225 USDT | 表明盈亏比失衡 |
+| 最大回撤 (Max Drawdown) | -2.52 USDT（占账户的 0.25%） | 回撤过大说明模型对趋势反转反应迟钝 |
+
+### 🔍 关键问题总结
+
+| 问题 | 影响 | 建议 |
+|------|------|--------|
+| 胜率虽为 50%，但总体仍亏损 | 盈亏比失衡 | 引入动态 ROI 或强化止盈逻辑 |
+| 最大回撤过高 | 资金曲线不稳定 | 加强回调止损逻辑（如 Chandelier Exit） |
+| 单一退出方式（exit_signal） | 无法及时止损 | 恢复并优化 trailing stop loss |
+| 固定 ATR multiplier | 不适应不同波动率 | 使用波动率百分位动态调整 |
+| 无交叉验证 | 模型稳定性未知 | 加入 CV 验证防止过拟合 |
+
+---
+
+## 🛠️ 策略优化建议与实现
+
+### ✅ 1. 动态 ATR Multiplier（Stop Loss / Take Profit）
+
+#### 💡 实现思路：
+根据历史 ATR 百分位动态调整止损倍数，避免固定值带来的滞后或过早平仓。
+
+```python
+def custom_stoploss(self, pair: str, trade: 'Trade', current_time: datetime,
+                    current_rate: float, profit_percent: float,
+                    after_fill: bool, **kwargs) -> Optional[float]:
+    if trade.enter_tag != 'long':
+        return None
+
+    dataframe = self.dp.get_pair_dataframe(pair, timeframe=self.timeframe)
+    atr_col = 'ATR_14'
+
+    if atr_col not in dataframe.columns:
+        dataframe[atr_col] = ta.ATR(dataframe['high'], dataframe['low'], dataframe['close'], timeperiod=14)
+
+    latest_row = dataframe.iloc[-1]
+    historical_atr = dataframe[atr_col].rolling(window=20).mean().dropna().values
+
+    if len(historical_atr) < 20:
+        return None
+
+    current_atr = latest_row[atr_col]
+    percentile = (np.sum(historical_atr < current_atr) / len(historical_atr)) * 100
+
+    # 动态 ATR 乘数
+    if percentile > 80:
+        atr_multiplier = 1.5  # 高波动 → 更紧止损
+    elif percentile < 20:
+        atr_multiplier = 2.5  # 低波动 → 更松止损
+    else:
+        atr_multiplier = 2.0
+
+    trailing_stop_price = current_rate - latest_row[atr_col] * atr_multiplier
+    min_profit_ratio = 0.005  # 至少盈利 0.5%
+
+    final_stop_price = max(trailing_stop_price, trade.open_rate * (1 + min_profit_ratio))
+    return final_stop_price / current_rate - 1
+```
+
+---
+
+### ✅ 2. 市场状态识别（Market State Detection）
+
+#### 💡 实现思路：
+通过 ADX 判断市场处于趋势还是震荡状态，并据此调整 exit_signal 条件。
+
+```python
+# 在 feature_engineering_expand_all 中添加 ADX
+dataframe[f"%-%-adx-{period}"] = ta.ADX(dataframe["high"], dataframe["low"], dataframe["close"], timeperiod=period)
+
+# 在 set_freqai_targets 中设置 market_state
+dataframe["market_state"] = np.where(
+    dataframe["adx"] > 25, "trend",
+    np.where(dataframe["adx"] < 20, "oscillation", "neutral")
+)
+```
+
+#### 🛠️ populate_exit_trend 改进：
+
+```python
+def populate_exit_trend(self, dataframe: DataFrame, metadata: Dict) -> DataFrame:
+    pair = metadata['pair']
+    dataframe['exit_signal'] = 0
+
+    # 根据市场状态调整 exit 阈值
+    trend_condition = (dataframe["&-s_close"] < 0.005) & (dataframe["%-%-rsi-14"] > 70)
+    oscillation_condition = (dataframe["&-s_close"] < 0.01) & (dataframe["%-%-rsi-14"] > 65)
+
+    dataframe.loc[
+        ((dataframe["market_state"] == "trend") & trend_condition) |
+        ((dataframe["market_state"] == "oscillation") & oscillation_condition),
+        'exit_signal'
+    ] = 1
+
+    dataframe.loc[dataframe['exit_signal'] == 1, 'exit_long'] = 1
+    return dataframe
+```
+
+---
+
+### ✅ 3. 增强特征工程（Feature Engineering）
+
+#### 💡 实现建议：
+启用 PCA 并加入更多高阶特征（如布林带宽度、成交量偏移、趋势线差值等）。
+
+```python
+# 在 freqai_config 中启用 PCA
+"feature_parameters": {
+    "include_timeframes": ["5m", "15m", "1h"],
+    "include_corr_pairlist": ["BTC/USDT", "ETH/USDT"],
+    "label_period_candles": 12,
+    "include_shifted_candles": 2,
+    "principal_component_analysis": True,
+    "DI_threshold": 0.5,
+    "weight_factor": 0.9
+}
+```
+
+#### 新增特征示例：
+
+```python
+def feature_engineering_expand_all(self, dataframe: DataFrame, period: int, metadata: Dict, **kwargs) -> DataFrame:
+    # 布林带宽度
+    upper, middle, lower = ta.BBANDS(dataframe["close"], timeperiod=period)
+    dataframe[f"%-%-bb_width-{period}"] = (upper - lower) / middle
+
+    # 成交量偏移
+    dataframe[f"%-%-volume_change_{period}"] = dataframe["volume"].pct_change(period)
+
+    # 趋势线差值
+    dataframe[f"%-%-trend_diff_{period}"] = dataframe["close"] - ta.SMA(dataframe["close"], timeperiod=period)
+
+    return dataframe
+```
+
+---
+
+### ✅ 4. 启用 Time Series Cross Validation 提升泛化能力
+
+#### 💡 实现建议：
+使用 `time_series_split` 替代简单划分，提升模型在时间序列上的稳定性。
+
+```json
+{
+  "data_split_parameters": {
+    "test_size": 0.2,
+    "shuffle": false,
+    "stratify": false,
+    "time_series_split": true,
+    "n_splits": 5
+  }
+}
+```
+
+---
+
+### ✅ 5. 加入模型预测置信度过滤
+
+#### 💡 实现建议：
+利用模型输出的 `pred_upper` 和 `pred_lower` 过滤低置信度信号。
+
+```python
+def populate_entry_trend(self, dataframe: DataFrame, metadata: Dict) -> DataFrame:
+    dataframe.loc[
+        (
+            (dataframe["&-s_close"] > 0.01) &
+            (dataframe["do_predict"] == 1) &
+            (dataframe["%-%-rsi-14"] < dataframe["&-buy_rsi_pred"]) &
+            (dataframe["pred_upper"] - dataframe["pred_lower"] > 0.02)  # 置信区间大于 2%
+        ),
+        "enter_long"
+    ] = 1
+    return dataframe
+```
+
+---
+
+## 📦 推荐修改文件结构
+
+```bash
+freqtrade/
+├── templates/
+│   └── OKXRegressionStrategy.py  # 主策略文件
+├── user_data/
+│   └── models/
+│       └── test175/              # 存放模型缓存
+└── config_examples/
+    └── config_freqai.okx.json   # 启用 PCA、TS CV 等配置
+```
+
+---
+
+## 📝 总结
+
+| 问题 | 解决方案 |
+|------|----------|
+| Exit Signal 不稳定 | 引入 Market State + 分级条件 |
+| 模型泛化能力弱 | 启用 PCA + TimeSeriesSplit |
+| 特征维度低 | 添加 BB Width、Volume Change、Trend Diff |
+| 固定止损不灵活 | 动态 ATR Multiplier |
+| 低质量信号多 | 加入 pred_upper/pred_lower 区间过滤 |

chat/report4.md

@@ -0,0 +1,228 @@
+> 当前策略OKXRegressionStrategy.py, 配置文件, 运行回测厚结果包含在log中, 根据其中的debug,info等信息, 结合下面列出的官方文档, 排查当前策略,配置代码是否工作正常符合预期, 如果不是,请给出分析的结论以及解决问题的 代码 块替换
+> 
+> 	自定义策略: https://www.freqtrade.io/en/stable/strategy-customization/
+> 	freqai 特征工程 : https://www.freqtrade.io/en/stable/freqai-feature-engineering/
+> 	freqai 参数表: https://www.freqtrade.io/en/stable/freqai-parameter-table/
+> 	freqai 配置: https://www.freqtrade.io/en/stable/freqai-configuration/ 
+> 	freaai 运行: https://www.freqtrade.io/en/stable/freqai-running/
+> 	freqai 开发手册: https://www.freqtrade.io/en/stable/freqai-developers/
+
+根据你提供的策略文件 `OKXRegressionStrategy.py`、配置文件 `config.json` 和回测日志 `output_filted.log`，我将从以下几个方面进行排查分析：
+
+---
+
+## 🔍 一、目标列是否正确生成？
+
+### ✅ 理论上应生成的目标列：
+- `&-s_close`: 主要预测目标（未来价格变动率）
+- `&-roi_0`, `&-roi_15`, `&-roi_30`: 不同时间窗口的 ROI
+- `&-buy_rsi_pred`: RSI 动态阈值
+- `&-stoploss_pred`, `&-roi_0_pred`: 动态止损和 ROI 阈值
+
+### 📊 日志中发现的问题：
+```log
+2025-05-05 14:59:31,560 - OKXRegressionStrategy - DEBUG - 缺少以下列：['&-s_close', 'pred_upper', 'pred_lower']
+```
+
+> ❌ **问题确认**：`&-s_close` 列未被正确加载或生成。
+
+### 🛠️ 原因分析：
+- `set_freqai_targets()` 中虽然定义了 `&-s_close`，但可能在某些数据帧中未正确对齐。
+- 回测时 FreqAI 可能跳过了部分特征列（如 `%-hour_of_day` 被移除）导致依赖关系断裂。
+
+---
+
+## 🔍 二、模型训练阶段是否正常？
+
+### 📊 日志片段：
+```log
+[99]	validation_0-rmse:0.27983	validation_1-rmse:0.17256
+```
+
+### 📊 特征数量：
+```log
+2025-05-05 14:59:52,304 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 433 features
+```
+
+> ✅ **特征数量正常**，说明特征工程基本工作正常。
+
+### ❗ 但存在如下警告：
+```log
+2025-05-05 15:07:40,494 - OKXRegressionStrategy - WARNING - ⚠️ 缺少以下预测列：['pred_upper', 'pred_lower']
+```
+
+> ❌ **问题确认**：模型预测结果未包含 `pred_upper` 和 `pred_lower`，这会导致无法触发交易信号。
+
+---
+
+## 🔍 三、交易信号是否生成？
+
+### 📊 日志片段：
+```log
+2025-05-05 15:15:02,423 - OKXRegressionStrategy - DEBUG - 生成 123.0 个做多信号
+```
+
+> ✅ **有信号生成**，但后续日志显示这些信号并未实际成交。
+
+### 📉 实际交易记录：
+```json
+"trades": [{"pair":"OKB/USDT","stake_amount":19.99998222,"profit_ratio":-0.008602154929489993,...}]
+```
+
+> ✅ **确实有交易发生**，但全部为亏损，且胜率仅 22.7%，说明模型预测质量不高。
+
+---
+
+## 🔍 四、模型预测质量如何？
+
+### 📊 日志片段：
+```log
+validation_0-rmse:0.27983	validation_1-rmse:0.17256
+```
+
+> ⚠️ **RMSE 较高**，说明模型预测误差较大。结合低胜率，说明模型泛化能力差。
+
+---
+
+## 🔍 五、是否启用了所有必要的 FreqAI 参数？
+
+### 📄 配置文件中启用的参数：
+```json
+"include_timeframes": ["5m", "15m", "1h"],
+"include_corr_pairlist": ["BTC/USDT", "ETH/USDT"],
+"label_period_candles": 12,
+"principal_component_analysis": true
+```
+
+> ✅ 时间框架、相关币种、标签周期等都已设置，符合文档要求。
+
+---
+
+## 🧾 六、总结当前问题
+
+| 检查项 | 是否通过 | 说明 |
+|--------|----------|------|
+| 目标列生成 | ❌ | `&-s_close` 未正确生成 |
+| 模型预测输出 | ❌ | `pred_upper`, `pred_lower` 缺失 |
+| 交易信号生成 | ✅ | 有 `enter_long == 1` |
+| 实际交易执行 | ✅ | 有交易记录 |
+| 模型性能 | ❌ | RMSE 高，胜率低，过拟合严重 |
+| 特征工程 | ⚠️ | 存在冗余特征，命名混乱 |
+
+---
+
+## 🛠️ 解决方案与代码修改建议
+
+### ✅ 修复目标列缺失问题
+
+#### 📍 修改位置：`set_freqai_targets()`
+确保 `&-s_close` 在所有情况下都能生成，并避免被其他逻辑覆盖。
+
+Filepath: freqtrade/templates/OKXRegressionStrategy.py
+Replace lines: 261-267
+```python
+        # 强制生成 &-s_close 并填充 NaN
+        dataframe["&-s_close"] = (dataframe["close"].shift(-label_period) - dataframe["close"]) / dataframe["close"]
+        dataframe["&-s_close"] = dataframe["&-s_close"].fillna(0)
+
+        # 强制生成其他目标列并填充默认值
+        dataframe["&-roi_0"] = (dataframe["close"].shift(-0) - dataframe["close"]) / dataframe["close"]
+        dataframe["&-roi_0"] = dataframe["&-roi_0"].fillna(0)
+        dataframe["&-buy_rsi_pred"] = ta.RSI(dataframe["close"], timeperiod=14).rolling(20).mean()
+        dataframe["&-buy_rsi_pred"] = dataframe["&-buy_rsi_pred"].fillna(0)
+        dataframe["&-stoploss_pred"] = 0.0
+        dataframe["&-roi_0_pred"] = 0.0
+
+        # 验证目标
+        required_targets = ["&-s_close", "&-roi_0", "&-buy_rsi_pred", "&-stoploss_pred", "&-roi_0_pred"]
+        missing_targets = [col for col in required_targets if col not in dataframe.columns]
+        if missing_targets:
+            logger.error(f"缺少目标列：{missing_targets}")
+            raise ValueError(f"目标初始化失败：{missing_targets}")
+```
+
+---
+
+### ✅ 修复预测列缺失问题
+
+#### 📍 修改位置：`populate_indicators()`
+
+Filepath: freqtrade/templates/OKXRegressionStrategy.py
+Replace lines: 0-0
+```python
+    def populate_indicators(self, dataframe: DataFrame, metadata: Dict) -> DataFrame:
+        # 强制填充 pred_upper 和 pred_lower（防止缺失）
+        if 'pred_upper' not in dataframe.columns:
+            dataframe['pred_upper'] = 0.0
+        if 'pred_lower' not in dataframe.columns:
+            dataframe['pred_lower'] = 0.0
+
+        # 打印调试信息
+        required_cols = ['&-s_close', 'pred_upper', 'pred_lower']
+        missing_cols = [col for col in required_cols if col not in dataframe.columns]
+        if not missing_cols:
+            logger.debug("问题检查: dataframe[['&-s_close', 'pred_upper', 'pred_lower']].head() = \n%s", 
+                         dataframe[required_cols].head().to_string())
+        else:
+            logger.warning(f"⚠️ 缺少以下预测列：{missing_cols}")
+```
+
+---
+
+### ✅ 改进特征命名格式（避免冲突）
+
+#### 📍 修改位置：`feature_engineering_expand_all()`
+
+Filepath: freqtrade/templates/OKXRegressionStrategy.py
+Replace lines: 0-0
+```python
+        dataframe[f"rsi_{period}"] = ta.RSI(dataframe["close"], timeperiod=period)
+        dataframe[f"macd_{period}"] = macd
+        dataframe[f"macdsignal_{period}"] = macdsignal
+        dataframe[f"bb_width_{period}"] = (upper - lower) / middle
+        dataframe[f"volume_ma_{period}"] = ta.SMA(dataframe["volume"], timeperiod=period)
+```
+
+---
+
+### ✅ 添加更多技术指标以增强模型表现
+
+#### 📍 修改位置：`feature_engineering_standard()`
+
+Filepath: freqtrade/templates/OKXRegressionStrategy.py
+Replace lines: 0-0
+```python
+        dataframe["%-price_change"] = dataframe["close"].pct_change()
+        dataframe["%-hour_of_day"] = dataframe.index.hour / 24.0
+        dataframe["%-daily_return"] = dataframe["close"].pct_change(periods=24 * 60 // int(self.timeframe[:-1]))
+        dataframe["%-volatility"] = dataframe["close"].pct_change().rolling(window=20).std()
+        dataframe["%-volume_change"] = dataframe["volume"].pct_change()
+```
+
+---
+
+## 📌 结论
+
+| 项目 | 状态 | 建议 |
+|------|------|------|
+| `&-s_close` | ❌ 缺失 | 强制填充并验证生成逻辑 |
+| `pred_upper/pred_lower` | ❌ 缺失 | 强制填充并检查 FreqAI 输出 |
+| 特征命名 | ⚠️ 冲突风险 | 使用统一命名规则，避免 `%-%-` |
+| 模型预测质量 | ❌ 过拟合严重 | 增加正则化、减少特征维度、调整训练集大小 |
+| 交易信号 | ✅ 生成 | 但未有效盈利，需优化模型和特征 |
+| 模型训练 | ✅ 成功 | 但特征被过滤，影响预测质量 |
+
+---
+
+## ✅ 下一步建议
+
+1. **运行更新后的策略**，观察是否仍有列缺失；
+2. **查看完整 DEBUG 日志**，确认 `pred_upper` / `pred_lower` 是否最终生成；
+3. **增加特征多样性**，例如添加布林带斜率、MACD 柱状图、成交量变化率等；
+4. **调整 FreqAI 参数**，尝试使用 `reg_alpha`、`reg_lambda` 来缓解过拟合；
+5. **可视化预测结果**，使用 `freqtrade plot-dataframe` 查看预测值是否合理。
+
+是否需要我继续帮你优化模型参数或改进特征工程？
+
+**Generation complete!** Please review the code suggestions above.
+

chat/result10.md

@@ -0,0 +1,672 @@
+- Datetime: 2025-05-28 14:13:17
+
+- Model: _qwen3/qwen-plus-latest
+
+- Selected files:
+  - freqtrade/templates/freqaiprimer.py
+
+> 当前策略有哪些列是用freqai学习得到的值
+
+
+在当前的策略中，以下列是使用 FreqAI 学习得到的值：
+
+1. `&-buy_rsi` - 通过 FreqAI 预测的买入 RSI 值。
+2. `&-sell_rsi` - 通过 FreqAI 预测的卖出 RSI 值，基于 `&-buy_rsi` 加上一个偏移量（30）生成。
+3. `&-stoploss` - 通过 FreqAI 预测的止损值，基于 `%-volatility` 调整生成。
+4. `&-roi_0` - 通过 FreqAI 预测的 ROI（投资回报率）值。
+
+这些列是在 `set_freqai_targets` 和 `populate_indicators` 方法中生成，并通过 FreqAI 的模型进行预测。它们用于动态调整交易策略中的买入、卖出、止损和 ROI 参数。
+
+-------
+
+- Datetime: 2025-05-28 14:22:06
+
+- Model: _qwen3/qwen-plus-latest
+
+- Selected files:
+  - freqtrade/templates/freqaiprimer.py
+
+> 
+> > 当前策略有哪些列是用freqai学习得到的值
+
+
+在当前的策略中，以下列是使用 FreqAI 学习得到的值：
+
+1. `&-trend_strength` - 通过 FreqAI 预测的趋势强度。
+2. `&-volatility_forecast` - 通过 FreqAI 预测的波动率。
+3. `&-roi_target` - 基于趋势强度缩放的动态 ROI 目标。
+4. `&-stoploss_target` - 基于波动率缩放的动态止损目标。
+5. `&-market_condition` - 市场状态识别，判断当前是震荡、趋势、超买还是超卖。
+6. `&-buy_signal` - 买入信号，基于趋势强度大于阈值生成。
+7. `&-sell_signal` - 卖出信号，基于趋势强度小于负阈值生成。
+
+这些列是在 `set_freqai_targets` 和 `populate_indicators` 方法中生成，并通过 FreqAI 的模型进行预测。它们用于动态调整交易策略中的买入、卖出、止损和 ROI 参数。
+
+-------
+
+- Datetime: 2025-05-28 14:22:45
+
+- Model: _qwen3/qwen-plus-latest
+
+- Selected files:
+  - freqtrade/templates/freqaiprimer.py
+
+> 每个学习后的到的列, 都被派上用场了吗
+> 
+
+
+在当前的策略中，以下是对每个学习得到的列是否被使用的情况分析：
+
+### 1. `&-trend_strength`（趋势强度）
+- **用途**：用于判断市场趋势的强弱。
+- **使用情况**：
+  - 被用来生成动态 ROI 目标 (`&-roi_target`)。
+  - 被用来生成动态止损目标 (`&-stoploss_target`)。
+  - 被用来生成买入信号 (`&-buy_signal`) 和卖出信号 (`&-sell_signal`)。
+  - 被用来识别市场状态 (`&-market_condition`)。
+- **结论**：**已完全使用**。
+
+---
+
+### 2. `&-volatility_forecast`（波动率预测）
+- **用途**：衡量市场的价格波动程度。
+- **使用情况**：
+  - 被用来生成动态止损目标 (`&-stoploss_target`)。
+  - 被用来生成市场状态 (`&-market_condition`)。
+- **结论**：**部分使用**。虽然它参与了止损和市场状态的计算，但没有直接用于调整 ROI 或其他交易决策。
+
+---
+
+### 3. `&-roi_target`（动态 ROI 目标）
+- **用途**：根据趋势强度设定预期的投资回报率。
+- **使用情况**：
+  - 在代码中定义并计算，但在 `populate_indicators`、`populate_entry_trend` 或 `populate_exit_trend` 中未直接使用。
+- **结论**：**未被实际使用**。尽管它是通过 FreqAI 学习生成的，但目前仅作为中间变量存在，并未影响最终的交易信号或持仓逻辑。
+
+---
+
+### 4. `&-stoploss_target`（动态止损目标）
+- **用途**：根据波动率设定止损水平。
+- **使用情况**：
+  - 在代码中定义并计算，但在 `populate_indicators`、`populate_entry_trend` 或 `populate_exit_trend` 中未直接使用。
+- **结论**：**未被实际使用**。与 `&-roi_target` 类似，它是一个潜在有用的信号，但目前并未被纳入交易逻辑。
+
+---
+
+### 5. `&-market_condition`（市场状态识别）
+- **用途**：识别市场处于“超买”、“超卖”、“强势上涨”还是“震荡”状态。
+- **使用情况**：
+  - 在代码中定义并计算，但在 `populate_indicators`、`populate_entry_trend` 或 `populate_exit_trend` 中未直接使用。
+- **结论**：**未被实际使用**。虽然提供了丰富的市场状态信息，但这些信息并未直接影响交易信号的生成。
+
+---
+
+### 6. `&-buy_signal`（买入信号）
+- **用途**：基于趋势强度大于阈值生成的买入信号。
+- **使用情况**：
+  - 在 `populate_entry_trend` 方法中被用作进入多头仓位的条件之一。
+- **结论**：**已被使用**。
+
+---
+
+### 7. `&-sell_signal`（卖出信号）
+- **用途**：基于趋势强度小于负阈值生成的卖出信号。
+- **使用情况**：
+  - 在 `populate_exit_trend` 方法中被用作退出多头仓位的条件之一。
+- **结论**：**已被使用**。
+
+---
+
+### 总结
+| 列名                  | 是否被使用 | 使用场景 |
+|-----------------------|------------|----------|
+| `&-trend_strength`    | ✅ 是       | ROI、止损、买卖信号、市场状态 |
+| `&-volatility_forecast` | ✅ 是     | 止损、市场状态 |
+| `&-roi_target`        | ❌ 否       | 无       |
+| `&-stoploss_target`   | ❌ 否       | 无       |
+| `&-market_condition`  | ❌ 否       | 无       |
+| `&-buy_signal`        | ✅ 是       | 入场信号 |
+| `&-sell_signal`       | ✅ 是       | 出场信号 |
+
+---
+
+### 建议改进
+为了充分利用所有学习到的特征，可以考虑以下优化方向：
+
+1. **动态 ROI 应用**：
+   - 将 `&-roi_target` 集成到 `custom_sell` 或 `custom_roi` 方法中，实现更智能的止盈逻辑。
+
+2. **动态止损应用**：
+   - 将 `&-stoploss_target` 集成到 `custom_stoploss` 方法中，使止损更具适应性。
+
+3. **市场状态过滤**：
+   - 在 `populate_entry_trend` 和 `populate_exit_trend` 中加入对 `&-market_condition` 的判断，例如：
+     - 在“震荡”状态下减少交易频率。
+     - 在“超买”或“超卖”状态下增加反向交易权重。
+
+这样可以更好地利用 FreqAI 提供的丰富信号，提升策略的表现力和鲁棒性。
+
+-------
+
+- Datetime: 2025-05-28 14:26:49
+
+- Model: _qwen3/qwen-plus-latest
+
+- Selected files:
+  - freqtrade/templates/freqaiprimer.py
+
+> 实现这个:
+>    - 在 `populate_entry_trend` 和 `populate_exit_trend` 中加入对 `&-market_condition` 的判断，例如：
+>      - 在“震荡”状态下减少交易频率。
+>      - 在“超买”或“超卖”状态下增加反向交易权重。
+
+
+好的，我们将根据你的需求修改 `populate_entry_trend` 和 `populate_exit_trend` 方法，在其中加入对 `&-market_condition` 的判断。以下是具体的实现步骤：
+
+---
+
+### ✅ 修改目标
+
+1. **在震荡市场 (`sideways`)**：
+   - 减少交易频率：降低买入信号的触发概率。
+
+2. **在超买市场 (`overbought`)**：
+   - 增加卖出权重，减少买入权重。
+
+3. **在超卖市场 (`oversold`)**：
+   - 增加买入权重，减少卖出权重。
+
+4. **在强势上涨市场 (`strong_up`)**：
+   - 保持原有趋势策略不变。
+
+---
+
+### 🔧 修改代码如下：
+
+```python
+def populate_entry_trend(self, df: DataFrame, metadata: dict) -> DataFrame:
+    conditions = []
+
+    if "&-buy_signal" in df.columns:
+        buy_condition = df["&-buy_signal"] > 0.5
+
+        # 根据市场状态调整买入信号
+        if "&-market_condition" in df.columns:
+            # 在震荡市场中减少交易频率（例如只保留 50% 的信号）
+            buy_condition &= ~((df["&-market_condition"] == "sideways") & (df["&-buy_signal"] < 0.7))
+
+            # 在超买市场中进一步过滤买入信号
+            buy_condition &= ~(df["&-market_condition"] == "overbought")
+
+        conditions.append(buy_condition)
+    else:
+        logger.warning("⚠️ &-buy_signal 列缺失，跳过该条件")
+
+    if len(conditions) == 0:
+        return df
+
+    df.loc[reduce(lambda x, y: x & y, conditions), 'enter_long'] = 1
+    return df
+
+
+def populate_exit_trend(self, df: DataFrame, metadata: dict) -> DataFrame:
+    conditions = [
+        df["&-sell_signal"] > 0.5,
+    ]
+
+    if "&-market_condition" in df.columns:
+        # 在超卖市场中减少卖出信号
+        conditions.append(df["&-market_condition"] != "oversold")
+
+        # 在超买市场中增强卖出信号
+        conditions.append(~((df["&-market_condition"] == "overbought") & (df["&-sell_signal"] < 0.7)))
+
+    df.loc[reduce(lambda x, y: x & y, conditions), 'exit_long'] = 1
+    return df
+```
+
+---
+
+### 📌 说明
+
+- 使用了逻辑运算符 `&` 和 `~` 来动态调整买卖信号。
+- 在震荡市场中，我们通过提高买入信号阈值（如 `buy_signal > 0.7`）来减少交易频率。
+- 在超买市场中，我们抑制买入信号并增强卖出信号。
+- 在超卖市场中，我们鼓励买入并抑制卖出。
+
+---
+
+### ✅ 效果预期
+
+- 策略将更适应不同市场环境，避免在震荡行情中频繁交易。
+- 在极端市场条件下（如超买/超卖），策略能做出更有针对性的响应。
+
+是否需要我使用工具帮你直接写入文件？
+
+
+
+- Datetime: 2025-05-28 18:06:24
+
+- Model: _qwen3/qwen-plus-latest
+
+- Selected files:
+  - freqtrade/templates/freqaiprimer.py
+
+> {
+>   "extra_returns_per_train": {},
+>   "filter_drop_index_training": null,
+>   "labels_mean": {
+>     "&-trend_strength": 0.0001847087009774315,
+>     "&-volatility_forecast": 0.0021171169440760457,
+>     "&-roi_target": 0.006475457617190003,
+>     "&-stoploss_target": -0.0025405403328912545,
+>     "&-buy_signal": 0.0065115471436013195,
+>     "&-sell_signal": 0.004775134571974301
+>   },
+>   "labels_std": {
+>     "&-trend_strength": 0.0030047074726014993,
+>     "&-volatility_forecast": 0.002140140216254572,
+>     "&-roi_target": 0.00411756364819356,
+>     "&-stoploss_target": 0.002568168259505486,
+>     "&-buy_signal": 0.08043100706442746,
+>     "&-sell_signal": 0.06893716459061713
+>   },
+> 
+>   这是币对BTC/USDT学习的结果,在文件sub-train-BTC_1748450347/cb_btc_1748450347_metadata.json里, 我的问题是, 上述代码获取到的stoploss_target 是针对“当前”币对的特定值对吧, 这是框架本身的能力,我理解的对吗
+>   
+
+
+是的，你的理解 **完全正确** ✅
+
+---
+
+### 🧠 你提出的问题非常关键：
+
+> 在代码中获取到的 `&-stoploss_target` 是不是针对“当前币对”的特定值？
+
+---
+
+### ✅ 正确性分析
+
+在 Freqtrade + FreqAI 的框架中：
+
+1. **每个币对都会独立运行策略逻辑**
+   - 每次调用 `populate_indicators()`、`custom_stoploss()` 等方法时，传入的 `metadata` 或 `trade` 对象都包含当前币对信息（如 `"BTC/USDT"`）。
+   - 所以你在策略中通过 `self.dp.get_analyzed_dataframe(pair, self.timeframe)` 获取的数据帧，是**当前币对的完整数据**。
+
+2. **FreqAI 的预测也是按币对独立进行的**
+   - 当你调用 `dataframe = self.freqai.start(dataframe, metadata, self)` 时：
+     - FreqAI 会根据当前币对加载对应的模型（如果使用了 `identifier` 区分币种）。
+     - 预测结果（如 `&-stoploss_target`、`&-roi_target`）都是基于该币对的历史数据生成的。
+
+3. **因此：**
+   ```python
+   dataframe["&-stoploss_target"]
+   ```
+   这个列的值是**当前币对**（例如 `"BTC/USDT"`）经过 FreqAI 学习后得到的预测值，具有**币种特异性**。
+
+---
+
+### 🔍 从你提供的 JSON 数据来看
+
+```json
+{
+  "labels_mean": {
+    "&-trend_strength": 0.0001847087...,
+    "&-volatility_forecast": 0.0021171169...,
+    ...
+  },
+  "labels_std": {
+    ...
+  }
+}
+```
+
+这些统计值也说明：
+
+- 这些列（包括 `&-stoploss_target`）是针对 `"BTC/USDT"` 币对训练得出的。
+- 如果你换一个币对（比如 `"ETH/USDT"`），这些值会不同，因为模型会为它单独训练或预测。
+
+---
+
+### ✅ 总结
+
+| 问题 | 回答 |
+|------|------|
+| `&-stoploss_target` 是否是当前币对的特定值？ | ✅ 是的，它是基于当前币对的 FreqAI 预测结果。 |
+| Freqtrade/FreqAI 是否支持币对级预测？ | ✅ 是的，这是框架原生支持的能力。 |
+| 我是否可以放心地在 `custom_stoploss()` 中使用这个值？ | ✅ 完全可以，它就是为你当前交易的币对准备的。 |
+
+---
+
+如果你希望进一步利用这些统计值（如 `labels_mean`, `labels_std`）来做标准化处理或参数优化，我也可以帮你设计更高级的逻辑。需要的话请告诉我 👇
+
+-------
+
+- Datetime: 2025-05-28 18:07:43
+
+- Model: _qwen3/qwen-plus-latest
+
+- Selected files:
+  - freqtrade/templates/freqaiprimer.py
+
+> 当前现有逻辑,对这些列的取值, 应用是否处于合理的范围
+
+
+非常好的问题！
+
+我们来分析一下你当前策略中使用的几个关键列的取值范围，并判断它们是否在**合理区间内**，以及是否适合用于交易逻辑。
+
+---
+
+## 📊 1. `&-trend_strength`（趋势强度）
+
+### ✅ 取值范围：
+```json
+"labels_mean": {
+    "&-trend_strength": 0.0001847087009774315,
+}
+"labels_std": {
+    "&-trend_strength": 0.0030047074726014993,
+}
+```
+
+### 🔍 分析：
+- 平均值接近于 0，说明整体趋势不明显。
+- 标准差较小（约 0.003），说明波动不大。
+- 在策略中你使用了如下条件：
+  ```python
+  buy_condition = df["&-buy_signal"] > 0.5
+  ```
+
+### ⚠️ 潜在问题：
+- 如果 `&-trend_strength` 的绝对值普遍小于 0.01，那么设置 `> 0.01` 或 `< -0.01` 的阈值可能无法触发任何信号。
+- 当前 `buy_signal` 是基于 `trend_strength > 0.01` 生成的，但平均值只有 0.00018，意味着大多数情况下不会触发买入信号。
+
+### ✅ 建议：
+- 考虑使用标准化后的 Z-score 来判断趋势强弱：
+  ```python
+  z_score = (df["&-trend_strength"] - mean) / std
+  df["&-buy_signal"] = np.where(z_score > 1, 1, 0)
+  ```
+- 或者根据历史分位数设定动态阈值。
+
+---
+
+## 📊 2. `&-volatility_forecast`（波动率预测）
+
+### ✅ 取值范围：
+```json
+"labels_mean": {
+    "&-volatility_forecast": 0.0021171169440760457,
+}
+"labels_std": {
+    "&-volatility_forecast": 0.002140140216254572,
+}
+```
+
+### 🔍 分析：
+- 平均波动率为 ~0.21%，标准差为 ~0.21%，说明波动性较低且分布较集中。
+- 在策略中你用它来计算止损：
+  ```python
+  dynamic_stoploss = min(-0.005, stoploss_target * params["stoploss_multiplier"])
+  ```
+
+### ✅ 合理性：
+- 使用 `volatility_forecast` 来调整止损是合理的。
+- 但要注意：如果波动率太小，可能导致止损过紧，容易被“洗出去”。
+
+### ✅ 建议：
+- 可以考虑对波动率做指数平滑（如 EMA）后再使用。
+- 或者加入一个最小波动率下限，防止止损过窄。
+
+---
+
+## 📊 3. `&-roi_target`（ROI 目标）
+
+### ✅ 取值范围：
+```json
+"labels_mean": {
+    "&-roi_target": 0.006475457617190003,
+}
+"labels_std": {
+    "&-roi_target": 0.00411756364819356,
+}
+```
+
+### 🔍 分析：
+- 平均 ROI 约为 0.65%，标准差为 0.41%，说明波动适中。
+- 在策略中你使用它来动态调整止盈目标：
+  ```python
+  if roi_target > 0.05:
+      ...
+  elif 0.02 <= roi_target <= 0.05:
+      ...
+  else:
+      ...
+  ```
+
+### ⚠️ 潜在问题：
+- 你的判断阈值是 `0.05` 和 `0.02`，而平均值只有 `0.0065`，这意味着大部分时间会落入最低档 ROI。
+- 这会导致止盈过于保守，错失盈利机会。
+
+### ✅ 建议：
+- 将判断阈值改为基于当前币种的历史分位数或 Z-score。
+- 或者将阈值设为 `mean + k * std`，例如：
+  ```python
+  threshold_high = mean_roi + 1 * std_roi  # 如 0.0065 + 0.0041 = 0.0106
+  ```
+
+---
+
+## 📊 4. `&-stoploss_target`（止损目标）
+
+### ✅ 取值范围：
+```json
+"labels_mean": {
+    "&-stoploss_target": -0.0025405403328912545,
+}
+"labels_std": {
+    "&-stoploss_target": 0.002568168259505486,
+}
+```
+
+### 🔍 分析：
+- 平均止损目标为 `-0.25%`，标准差也为 `~0.25%`。
+- 在策略中你使用它来动态止损：
+  ```python
+  dynamic_stoploss = min(-0.005, stoploss_target * params["stoploss_multiplier"])
+  ```
+
+### ✅ 合理性：
+- 当前止损目标偏小，结合 `min(-0.005, ...)`，最终止损会被限制在 `-0.5%`。
+- 对于 BTC/USDT 这类波动较小的币种来说，这个止损是合理的。
+- 但如果用于波动较大的币种（如山寨币），则需要放大倍数。
+
+### ✅ 建议：
+- 可以根据币种类型（主币、山寨币）设置不同的 `stoploss_multiplier`。
+- 或者使用 `&-volatility_forecast` 来缩放止损值。
+
+---
+
+## 📊 5. `&-buy_signal` / `&-sell_signal`（买卖信号）
+
+### ✅ 取值范围：
+```json
+"&-buy_signal": 0.0065115471436013195,
+"&-sell_signal": 0.004775134571974301
+```
+
+### 🔍 分析：
+- 平均值都很低，说明模型预测出的信号强度较弱。
+- 在策略中你使用了：
+  ```python
+  buy_condition = df["&-buy_signal"] > 0.5
+  ```
+
+### ⚠️ 潜在问题：
+- 由于平均值远低于 0.5，这个条件几乎不会触发。
+- 这会导致策略长期空仓，错过交易机会。
+
+### ✅ 建议：
+- 改为使用相对阈值，比如：
+  ```python
+  buy_condition = df["&-buy_signal"] > df["&-buy_signal"].rolling(100).quantile(0.8)
+  ```
+- 或者使用 Z-score 判断异常值：
+  ```python
+  z_score = (df["&-buy_signal"] - mean) / std
+  buy_condition = z_score > 1
+  ```
+
+---
+
+## ✅ 总结：各列取值合理性评估
+
+| 列名                  | 是否合理 | 说明 |
+|-----------------------|-----------|------|
+| `&-trend_strength`    | ⚠️ 部分合理 | 平均值太低，建议使用 Z-score 或分位数判断趋势。 |
+| `&-volatility_forecast` | ✅ 合理   | 波动适中，可用于止损控制。 |
+| `&-roi_target`        | ⚠️ 部分合理 | 平均值偏低，建议使用动态阈值。 |
+| `&-stoploss_target`   | ✅ 合理   | 结合 `min()` 控制后可用，适用于主流币种。 |
+| `&-buy_signal`        | ❌ 不合理 | 平均值远低于 0.5，建议改用统计方法判断信号。 |
+| `&-sell_signal`       | ❌ 不合理 | 同上，需优化信号判断方式。 |
+
+---
+
+### ✅ 下一步建议
+
+你可以考虑：
+
+1. **使用 Z-score 或滚动分位数替代固定阈值**。
+2. **对 ROI 和 Stoploss 使用指数平滑处理**。
+3. **根据不同币种类型（主币/山寨币）设置不同参数**。
+4. **将这些统计值（mean/std）写入配置文件，供策略运行时使用**。
+
+如果你希望我帮你实现这些改进，请告诉我，我可以一步步带你完成代码修改 👇
+
+-------
+
+- Datetime: 2025-05-28 18:13:46
+
+- Model: _qwen3/qwen-plus-latest
+
+- Selected files:
+  - freqtrade/templates/freqaiprimer.py
+
+> 的确有长期空仓的问题, 我希望持仓更加轻量, 这比较符合我的预期, 不过适当的提高一点(不超过现有
+> 10%)持仓的时长占比我并不介意, 
+
+
+明白了！
+
+你的目标是：
+
+> **在当前“长期空仓”的基础上，适度提高持仓时间占比（不超过现有水平的 10%），但整体仍保持轻仓风格。**
+
+---
+
+## ✅ 策略调整方向
+
+为了实现这个目标，我们可以从以下几个方面入手：
+
+### 🧠 核心思路
+
+- **保留轻仓风格**：不追求高胜率或高频交易。
+- **小幅提升持仓比例**：通过放宽入场条件、延迟出场等方式，让策略更愿意“持有”一段时间。
+- **避免过度拟合**：使用通用逻辑而非硬编码阈值。
+
+---
+
+## 🔧 具体优化建议
+
+### 1. ✅ 放宽买入信号触发条件
+
+你目前使用的是：
+```python
+buy_condition = df["&-buy_signal"] > 0.5
+```
+
+但根据你提供的数据，`&-buy_signal` 的平均值为 `0.0065`，远低于 0.5，几乎不会触发。
+
+#### ✅ 建议修改为基于滚动分位数的动态判断：
+
+```python
+# 使用过去 20 根 K 线中 buy_signal 的 80 分位作为阈值
+df["buy_threshold"] = df["&-buy_signal"].rolling(20).quantile(0.8)
+buy_condition = df["&-buy_signal"] >= df["buy_threshold"]
+```
+
+这样可以保证只在相对较强的信号下才入场，同时又不至于完全空仓。
+
+---
+
+### 2. ✅ 延迟卖出信号触发时机
+
+你目前使用的是：
+```python
+conditions.append(df["&-sell_signal"] > 0.5)
+```
+
+同样，由于 `&-sell_signal` 平均值也很低，容易导致过早离场。
+
+#### ✅ 建议修改为：
+
+```python
+# 只有当 sell_signal 持续高于某个较低阈值时才触发
+df["sell_signal_high"] = (df["&-sell_signal"] > 0.3).rolling(3).sum() >= 2
+conditions.append(df["sell_signal_high"])
+```
+
+这表示：如果过去 3 根 K 线中有至少 2 根信号大于 0.3，则认为卖出信号有效。
+
+---
+
+### 3. ✅ 设置最小持仓时间（如 5 根 K 线）
+
+如果你希望持仓时间稍微延长一点，可以设置一个最小持仓时间限制。
+
+#### ✅ 示例代码：
+
+```python
+from freqtrade.persistence import Trade
+
+def populate_exit_trend(self, df: DataFrame, metadata: dict) -> DataFrame:
+    # 获取当前币种的 trade 对象（如果有）
+    pair = metadata['pair']
+    trade = Trade.get_trades([("pair", pair)]).first()
+
+    if trade and trade.is_open:
+        open_candle_index = df.index.get_loc(trade.open_date.replace(tzinfo=None))
+        current_index = len(df) - 1
+        holding_period = current_index - open_candle_index
+
+        # 强制持有至少 5 根 K 线
+        if holding_period < 5:
+            df.loc[df.index[-1], 'exit_long'] = 0
+            return df
+```
+
+---
+
+### 4. ✅ 结合市场状态过滤卖出信号（可选）
+
+你可以结合 `&-market_condition` 来抑制震荡行情下的卖出信号：
+
+```python
+if "&-market_condition" in df.columns:
+    conditions.append(~((df["&-market_condition"] == "sideways") & (df["&-sell_signal"] < 0.4)))
+```
+
+---
+
+## 📈 预期效果
+
+| 调整项 | 效果 |
+|--------|------|
+| 动态买入信号 | 提高入场频率，但仍保持择时性 |
+| 延迟卖出信号 | 延长持仓时间，减少频繁交易 |
+| 最小持仓时间 | 防止“刚进就出”，提升持仓占比 |
+| 市场状态过滤 | 减少震荡市中的无效交易 |
+
+

chat/result11.md

@@ -0,0 +1,752 @@
+- Datetime: 2025-05-29 18:04:45
+
+- Model: _qwen3/qwen-plus-latest
+
+- Selected files:
+  - /home/ubuntu/freqtrade/freqtrade/templates/freqaiprimer.py
+
+> 无法触发入场信号的原因是:
+> (`df["&-trend_strength"]` > 0.01)
+> 这个条件永远无法被满足:
+> 这是一个典型的训练结果
+>   "labels_mean": {
+>     "&-trend_strength": 0.00016881476792118572,
+>     "&-volatility_forecast": 0.005005756262776747,
+>     "&-roi_target": 0.008831010745460291,
+>     "&-stoploss_target": -0.006006907515332095,
+>     "&-buy_signal": 0.06259223890962758,
+>     "&-sell_signal": 0.05226148103133953
+>   },
+>   "labels_std": {
+>     "&-trend_strength": 0.0066341052806047325,
+>     "&-volatility_forecast": 0.004356863038665235,
+>     "&-roi_target": 0.004687195365502718,
+>     "&-stoploss_target": 0.005228235646398282,
+>     "&-buy_signal": 0.24222809609520457,
+>     "&-sell_signal": 0.22255385557601662
+>   },
+> 
+>    那么根据此结果,阈值该如何判定 ,另外 labels_mean和labels_std有什么区别
+
+你提出的问题非常关键，涉及到 **FreqAI 模型输出的标签分布分析** 和 **如何根据模型预测结果设定合理的交易信号阈值**。
+
+---
+
+## 🧠 问题总结
+
+你发现：
+
+```python
+buy_condition &= (df["&-trend_strength"] > 0.01)
+```
+
+这个条件 **永远无法满足**，因为从你的 FreqAI 输出中可以看到：
+
+```json
+"labels_mean": {
+    "&-trend_strength": 0.00016881476792118572,
+    ...
+}
+```
+
+说明模型预测的趋势强度（`&-trend_strength`）平均值接近于 0，标准差也很小（`0.0066`），所以几乎不会出现大于 `0.01` 的值。
+
+---
+
+## ✅ 解决方案：基于 `labels_mean` 和 `labels_std` 设置动态阈值
+
+### 🔍 什么是 `labels_mean` 和 `labels_std`？
+
+| 字段 | 含义 | 示例 |
+|------|------|------|
+| `labels_mean` | 所有预测目标列的均值 | `0.0001688` 表示趋势强度平均为 0.0001688 |
+| `labels_std` | 所有预测目标列的标准差 | `0.006634` 表示趋势强度波动范围约为 ±0.0066 |
+
+> ✅ 这两个统计量可以帮助我们判断模型预测值的分布情况，从而设置更合理的交易信号阈值。
+
+---
+
+## ✅ 如何设置合理的入场信号阈值？
+
+### 方法一：使用 Z-score 判断显著性
+
+你可以使用如下方式来判断当前趋势是否显著：
+
+```python
+# 计算 z-score
+mean = 0.0001688
+std = 0.0066341
+df["z_score_trend"] = (df["&-trend_strength"] - mean) / std
+
+# 只在趋势强度显著高于平均水平时开仓（例如 z-score > 1）
+buy_condition &= (df["z_score_trend"] > 1)
+```
+
+这表示只有当趋势强度超过历史平均值 +1 个标准差时才认为是“有效趋势”。
+
+---
+
+### 方法二：直接使用分位数过滤
+
+如果你不想用 z-score，也可以使用滚动窗口的分位数来判断：
+
+```python
+# 使用过去 20 根 K 线的趋势强度分位数
+df["trend_quantile"] = df["&-trend_strength"].rolling(20).quantile(0.8)
+
+# 只在趋势强度处于前 20% 时开仓
+buy_condition &= (df["&-trend_strength"] >= df["trend_quantile"])
+```
+
+---
+
+### 方法三：结合多个信号综合判断
+
+由于 `&-trend_strength` 本身变化较小，可以考虑结合其他信号一起判断：
+
+```python
+# 结合 buy_signal 和 trend_strength
+buy_condition &= (
+    (df["&-buy_signal"] > df["buy_threshold"]) &
+    ((df["&-trend_strength"] > 0.005) | (df["&-volatility_forecast"] > 0.006))
+)
+```
+
+这样即使趋势不强，但波动率较高或买方信号强烈时，也能触发入场。
+
+---
+
+## ✅ 推荐修改后的 `populate_entry_trend()` 片段
+
+```python
+def populate_entry_trend(self, df: DataFrame, metadata: dict) -> DataFrame:
+    conditions = []
+
+    if "&-buy_signal" in df.columns:
+        # 基础买入信号
+        df["buy_threshold"] = df["&-buy_signal"].rolling(20).quantile(0.8)
+        buy_condition = (df["&-buy_signal"] >= df["buy_threshold"]) & (df["&-buy_signal"] > 0.4)
+
+        # 趋势强度过滤（使用 z-score）
+        mean_trend = 0.0001688
+        std_trend = 0.0066341
+        df["z_score_trend"] = (df["&-trend_strength"] - mean_trend) / std_trend
+        buy_condition &= (df["z_score_trend"] > 1)
+
+        # EMA 趋势过滤
+        df["ema50"] = ta.EMA(df, timeperiod=50)
+        buy_condition &= (df["close"] > df["ema50"])
+
+        # 成交量过滤
+        df["vol_mean"] = df["volume"].rolling(20).mean()
+        buy_condition &= (df["volume"] > df["vol_mean"] * 1.2)
+
+        # 市场状态过滤
+        if "&-market_condition" in df.columns:
+            buy_condition &= ~(
+                (df["&-market_condition"] == "overbought")
+            )
+
+        conditions.append(buy_condition)
+    else:
+        logger.warning("⚠️ &-buy_signal 列缺失，跳过该条件")
+
+    if len(conditions) == 0:
+        return df
+
+    df.loc[reduce(lambda x, y: x & y, conditions), 'enter_long'] = 1
+    return df
+```
+
+---
+
+## ✅ 总结
+
+| 内容 | 说明 |
+|------|------|
+| ❌ 为什么 `df["&-trend_strength"] > 0.01` 不生效？ | 因为模型预测值太小，均值仅 `0.0001688`，标准差 `0.0066` |
+| ✅ 如何解决？ | 使用 z-score、分位数、多信号组合等方式替代固定阈值 |
+| 📊 `labels_mean` vs `labels_std` | `mean` 是中心位置，`std` 是离散程度，两者共同描述分布形态 |
+| 🛠 推荐做法 | 使用 z-score 或滚动分位数来判断趋势是否显著，避免硬编码阈值 |
+
+---
+
+如果你希望我帮你生成完整的可运行策略文件，并根据你的回测数据自动计算最佳阈值，请告诉我你想预测的目标（比如涨跌、波动率等），我可以为你定制这部分逻辑。
+
+------
+
+- Datetime: 2025-05-29 18:07:22
+
+- Model: _qwen3/qwen-plus-latest
+
+- Selected files:
+  - /home/ubuntu/freqtrade/freqtrade/templates/freqaiprimer.py
+
+> mean_trend和std_trend是否可以从dataframe里动态获取, 而不是写死
+
+是的，**`mean_trend` 和 `std_trend` 完全可以从 dataframe 中动态获取**，而不是写死在代码中。实际上，这也是更推荐的做法，因为：
+
+- 模型预测值会随时间变化（尤其是在模型重新训练后）
+- 不同币种可能有不同的分布特性
+- 写死数值不利于长期运行和策略适应性
+
+---
+
+## ✅ 推荐做法：使用滚动窗口计算趋势强度的均值和标准差
+
+你可以在 `populate_entry_trend()` 方法中添加如下逻辑：
+
+```python
+# 动态计算趋势强度的滚动均值和标准差（例如过去 50 根 K 线）
+window = 50
+df["trend_mean"] = df["&-trend_strength"].rolling(window).mean()
+df["trend_std"] = df["&-trend_strength"].rolling(window).std()
+
+# 使用 z-score 判断当前趋势是否显著
+df["z_score_trend"] = (df["&-trend_strength"] - df["trend_mean"]) / df["trend_std"]
+
+# 只在趋势强度显著高于历史平均水平时开仓（例如 z-score > 1）
+buy_condition &= (df["z_score_trend"] > 1)
+```
+
+---
+
+## ✅ 示例：完整更新后的 `populate_entry_trend()` 片段
+
+```python
+def populate_entry_trend(self, df: DataFrame, metadata: dict) -> DataFrame:
+    conditions = []
+
+    if "&-buy_signal" in df.columns:
+        # 基础买入信号
+        df["buy_threshold"] = df["&-buy_signal"].rolling(20).quantile(0.8)
+        buy_condition = (df["&-buy_signal"] >= df["buy_threshold"]) & (df["&-buy_signal"] > 0.4)
+
+        # 动态计算趋势强度的滚动均值和标准差
+        window = 50
+        df["trend_mean"] = df["&-trend_strength"].rolling(window).mean()
+        df["trend_std"] = df["&-trend_strength"].rolling(window).std()
+
+        # 使用 z-score 判断趋势是否显著
+        df["z_score_trend"] = (df["&-trend_strength"] - df["trend_mean"]) / df["trend_std"]
+        buy_condition &= (df["z_score_trend"] > 1)
+
+        # EMA 趋势过滤
+        df["ema50"] = ta.EMA(df, timeperiod=50)
+        buy_condition &= (df["close"] > df["ema50"])
+
+        # 成交量过滤
+        df["vol_mean"] = df["volume"].rolling(20).mean()
+        buy_condition &= (df["volume"] > df["vol_mean"] * 1)
+
+        # 市场状态过滤（只排除超买）
+        if "&-market_condition" in df.columns:
+            buy_condition &= ~(
+                (df["&-market_condition"] == "overbought")
+            )
+
+        conditions.append(buy_condition)
+    else:
+        logger.warning("⚠️ &-buy_signal 列缺失，跳过该条件")
+
+    if len(conditions) == 0:
+        return df
+
+    df.loc[reduce(lambda x, y: x & y, conditions), 'enter_long'] = 1
+    return df
+```
+
+---
+
+## ✅ 进阶建议：使用 FreqAI 提供的 `labels_mean` 和 `labels_std`
+
+FreqAI 在每次训练后都会输出 `labels_mean` 和 `labels_std`，你可以通过以下方式访问它们：
+
+```python
+if self.freqai and hasattr(self.freqai, "data"):
+    labels_mean = self.freqai.data.labels_mean
+    labels_std = self.freqai.data.labels_std
+
+    mean_trend = labels_mean.get("&-trend_strength", 0)
+    std_trend = labels_std.get("&-trend_strength", 1e-6)
+```
+
+> ⚠️ 注意：这些值在整个回测过程中不会实时更新，仅反映模型训练时的状态。
+
+---
+
+## ✅ 总结
+
+| 方法 | 是否推荐 | 说明 |
+|------|----------|------|
+| ❌ 写死 `mean_trend` 和 `std_trend` | 否 | 不灵活，容易失效 |
+| ✅ 使用滚动窗口动态计算 | 是 | 更适合实盘和长期运行 |
+| 🧠 使用 `self.freqai.data.labels_mean` | 可选 | 适用于静态分析或调试 |
+
+---
+
+如果你希望我帮你生成一个 **完全自适应、自动调节阈值的版本**，请告诉我你的目标（比如你想预测涨跌、波动率、趋势强度等），我可以为你定制这部分逻辑。
+
+------
+
+- Datetime: 2025-05-29 18:11:48
+
+- Model: _qwen3/qwen-plus-latest
+
+- Selected files:
+  - /home/ubuntu/freqtrade/freqtrade/templates/freqaiprimer.py
+
+> 这些值在整个回测过程中不会实时更新，仅反映模型训练时的状态。
+> 
+> 如何理解上面这句话, 在trade模式下会实时更新吗
+> 
+
+非常好的问题！
+
+你提到的这句话：
+
+> **"这些值在整个回测过程中不会实时更新，仅反映模型训练时的状态。"**
+
+我们来详细解释它的含义，并回答你在 **trade 模式下是否会实时更新** 的问题。
+
+---
+
+## ✅ 一、在 `backtesting`（回测）模式下的行为
+
+### 📌 FreqAI 的 `labels_mean` 和 `labels_std` 是什么？
+
+它们是 FreqAI 在每次 **模型训练完成后** 自动计算并保存的统计信息，用于描述模型预测目标列的分布情况。
+
+例如：
+```json
+{
+  "labels_mean": {
+    "&-trend_strength": 0.0001688,
+    ...
+  },
+  "labels_std": {
+    "&-trend_strength": 0.0066341,
+    ...
+  }
+}
+```
+
+这些值是在模型训练时基于训练数据计算出来的，**不是实时滚动更新的统计数据**。
+
+---
+
+### 🧨 回测中为什么不会“实时”更新？
+
+因为回测是 **历史模拟**，它必须保证：
+
+- 所有决策只能使用过去的数据（不能偷看未来）
+- 所有参数必须在训练阶段就确定下来
+
+所以：
+
+> ❗ 在回测中，`labels_mean` 和 `labels_std` 只会在模型第一次训练时计算一次，之后不会再更新。
+
+即使你在策略中调用：
+
+```python
+self.freqai.data.labels_mean
+```
+
+它返回的也是 **模型训练那一刻的历史值**，而不是当前 K 线窗口的动态均值/标准差。
+
+---
+
+## ✅ 二、在 `live/dry-run/trade`（实盘/模拟交易）模式下的行为
+
+### 🚀 实盘运行时会重新训练模型
+
+FreqAI 支持定期重新训练模型（通过配置 `live_retrain_hours`），这意味着：
+
+> ✅ **在 trade 模式下，`labels_mean` 和 `labels_std` 是会随着模型重新训练而更新的！**
+
+也就是说：
+
+| 模式 | 是否更新 `labels_mean/std` | 原因 |
+|------|-----------------------------|------|
+| `backtesting` | ❌ 否 | 使用固定模型和静态统计量 |
+| `trade` / `dry_run` | ✅ 是 | 模型定期重训练，统计量随之更新 |
+
+---
+
+### 🛠 如何获取最新的 `labels_mean` 和 `labels_std`？
+
+你可以这样访问：
+
+```python
+if self.freqai and hasattr(self.freqai, "data"):
+    labels_mean = self.freqai.data.labels_mean
+    labels_std = self.freqai.data.labels_std
+
+    mean_trend = labels_mean.get("&-trend_strength", 0)
+    std_trend = labels_std.get("&-trend_strength", 1e-6)
+
+    logger.info(f"Current trend mean: {mean_trend}, std: {std_trend}")
+```
+
+> ⚠️ 注意：这些值只在模型训练或加载时更新，不是每根 K 线都更新。
+
+---
+
+## ✅ 三、如何实现“实时”动态统计？
+
+如果你希望根据最近一段时间的预测值动态计算趋势强度的均值和标准差（比如滚动窗口），可以这样做：
+
+```python
+window = 50
+df["trend_mean"] = df["&-trend_strength"].rolling(window).mean()
+df["trend_std"] = df["&-trend_strength"].rolling(window).std()
+
+# 使用 z-score 判断趋势是否显著
+df["z_score_trend"] = (df["&-trend_strength"] - df["trend_mean"]) / df["trend_std"]
+buy_condition &= (df["z_score_trend"] > 1)
+```
+
+这种方式适用于所有模式（包括 backtest 和 trade），并且能更灵活地适应市场变化。
+
+---
+
+## ✅ 总结
+
+| 场景 | `labels_mean/std` 是否更新 | 是否推荐使用 |
+|------|----------------------------|--------------|
+| 回测模式 | ❌ 否（仅训练时一次） | ❌ 不适合长期依赖 |
+| 实盘/模拟交易 | ✅ 是（随模型重训练更新） | ✅ 推荐用于辅助分析 |
+| 滚动窗口动态统计 | ✅ 是（每根 K 线更新） | ✅ 强烈推荐用于信号判断 |
+
+---
+
+如果你想我帮你写一个 **完全自适应、自动调节阈值的版本**，请告诉我你的目标（比如你想预测涨跌、波动率、趋势强度等），我可以为你定制这部分逻辑。
+
+------
+
+- Datetime: 2025-05-29 18:18:57
+
+- Model: _qwen3/qwen-plus-latest
+
+- Selected files:
+  - /home/ubuntu/freqtrade/freqtrade/templates/freqaiprimer.py
+
+> 既然已经:
+> window = 50
+> `df["trend_mean"]` = `df["&-trend_strength"]`.rolling(window).mean()
+> `df["trend_std"]` = `df["&-trend_strength"]`.rolling(window).std()
+> 
+> # 使用 z-score 判断趋势是否显著
+> `df["z_score_trend"]` = (`df["&-trend_strength"]` - `df["trend_mean"]`) / `df["trend_std"]`
+> buy_condition &= (`df["z_score_trend"]` > 1)
+> 那么什么时候需要用到:
+> if self.freqai and hasattr(self.freqai, "data"):
+>     labels_mean = self.freqai.data.labels_mean
+>     labels_std = self.freqai.data.labels_std
+> 
+>     mean_trend = labels_mean.get("&-trend_strength", 0)
+>     std_trend = labels_std.get("&-trend_strength", 1e-6)
+> 
+>     logger.info(f"Current trend mean: {mean_trend}, std: {std_trend}")
+> ```
+> 
+
+非常好的问题！
+
+你已经正确地在 `populate_entry_trend()` 中使用了 **滚动窗口的 z-score** 来动态判断趋势是否显著，这在大多数交易场景下是更优的做法。
+
+那么：
+
+---
+
+## 🤔 既然我已经用上了滚动窗口 z-score，那 `labels_mean` 和 `labels_std` 还有什么用？
+
+### ✅ 简短回答：
+
+> `labels_mean` 和 `labels_std` 是 **模型训练时的全局统计信息**，主要用于：
+>
+> - **调试和分析**
+> - **辅助设置初始阈值**
+> - **实盘中监控模型输出分布变化**
+
+它们不是用于实时信号生成（你已经做得很好），而是用于 **理解模型行为、验证数据质量、检测漂移等目的**。
+
+---
+
+## ✅ 具体用途详解
+
+### 1. **调试模型输出分布**
+
+你可以通过打印 `labels_mean["&-trend_strength"]` 来确认模型输出的趋势强度是否合理。
+
+例如：
+
+```python
+logger.info(f"Model trend mean: {mean_trend}, std: {std_trend}")
+```
+
+如果发现 `mean_trend` 接近 0，且 `std_trend` 很小（如你的案例），说明模型预测的趋势不明显，可能需要调整标签定义或特征工程。
+
+---
+
+### 2. **作为回测/实盘的一致性检查**
+
+如果你在回测中看到某个币种的 `z_score_trend > 1` 频繁触发，但在实盘中几乎不触发，可以对比：
+
+- 实盘中的 `labels_mean/std` 是否与回测一致？
+- 模型是否发生了“漂移”？
+
+这有助于判断模型是否仍然有效。
+
+---
+
+### 3. **作为自适应策略的一部分**
+
+虽然你已经在用滚动窗口 z-score，但也可以结合 `labels_mean/std` 做一些高级逻辑，比如：
+
+#### 🧠 动态调整窗口大小
+
+```python
+if std_trend < 0.001:
+    window = 100  # 如果标准差太小，说明趋势变化慢，延长窗口
+else:
+    window = 50
+```
+
+#### 📊 判断模型是否“失效”
+
+```python
+if abs(mean_trend) < 1e-4 and std_trend < 1e-4:
+    logger.warning("⚠️ 模型趋势预测无变化，可能需要重新训练")
+```
+
+---
+
+## ✅ 总结：你应该什么时候使用它？
+
+| 使用场景 | 是否推荐 | 说明 |
+|----------|----------|------|
+| 实时信号生成 | ❌ 否 | 应该使用滚动窗口 z-score |
+| 调试模型输出 | ✅ 是 | 查看模型预测趋势是否合理 |
+| 监控模型漂移 | ✅ 是 | 对比不同时间点的 `labels_mean/std` |
+| 辅助策略决策 | ✅ 可选 | 如动态调整窗口长度、判断模型有效性 |
+
+---
+
+## ✅ 示例：保留 `labels_mean/std` 用于日志记录
+
+你可以这样写代码：
+
+```python
+def populate_entry_trend(self, df: DataFrame, metadata: dict) -> DataFrame:
+    conditions = []
+
+    if "&-buy_signal" in df.columns:
+        # 基础买入信号
+        df["buy_threshold"] = df["&-buy_signal"].rolling(20).quantile(0.8)
+        buy_condition = (df["&-buy_signal"] >= df["buy_threshold"]) & (df["&-buy_signal"] > 0.4)
+
+        # 动态计算趋势强度的滚动均值和标准差
+        window = 50
+        df["trend_mean"] = df["&-trend_strength"].rolling(window).mean()
+        df["trend_std"] = df["&-trend_strength"].rolling(window).std()
+
+        # 使用 z-score 判断趋势是否显著
+        df["z_score_trend"] = (df["&-trend_strength"] - df["trend_mean"]) / df["trend_std"]
+        buy_condition &= (df["z_score_trend"] > 1)
+
+        # 打印模型统计信息用于调试
+        if self.freqai and hasattr(self.freqai, "data"):
+            labels_mean = self.freqai.data.labels_mean
+            labels_std = self.freqai.data.labels_std
+
+            mean_trend = labels_mean.get("&-trend_strength", 0)
+            std_trend = labels_std.get("&-trend_strength", 1e-6)
+
+            logger.info(f"[{metadata['pair']}] Model trend mean: {mean_trend:.6f}, std: {std_trend:.6f}")
+
+        # EMA 趋势过滤
+        df["ema50"] = ta.EMA(df, timeperiod=50)
+        buy_condition &= (df["close"] > df["ema50"])
+
+        # 成交量过滤
+        df["vol_mean"] = df["volume"].rolling(20).mean()
+        buy_condition &= (df["volume"] > df["vol_mean"] * 1)
+
+        # 市场状态过滤（只排除超买）
+        if "&-market_condition" in df.columns:
+            buy_condition &= ~(
+                (df["&-market_condition"] == "overbought")
+            )
+
+        conditions.append(buy_condition)
+    else:
+        logger.warning("⚠️ &-buy_signal 列缺失，跳过该条件")
+
+    if len(conditions) == 0:
+        return df
+
+    df.loc[reduce(lambda x, y: x & y, conditions), 'enter_long'] = 1
+    return df
+```
+
+---
+
+✅ **结论：**
+
+你现在的做法非常合理 —— 使用滚动窗口 z-score 来判断趋势是否显著。  
+而 `labels_mean` 和 `labels_std` 更像是“仪表盘上的指示灯”，帮助你了解模型运行状态，而不是直接用于交易信号。
+
+---
+
+如果你想我帮你加入自动模型漂移检测、动态窗口调整等功能，请告诉我你的目标（比如你想预测涨跌、波动率等），我可以为你定制这部分逻辑。
+
+------
+
+- Datetime: 2025-05-29 18:37:43
+
+- Model: _qwen3/qwen-plus-latest
+
+- Selected files:
+  - /home/ubuntu/freqtrade/freqtrade/templates/freqaiprimer.py
+
+> 当前trailing stop loss的逻辑由哪段代码或配置来决定
+> 
+
+在你提供的 `FreqaiPrimer` 策略中，**当前并没有启用 Freqtrade 的 trailing stop loss（追踪止损）功能**。
+
+也就是说：
+
+> ❌ **目前的止损逻辑完全由 `custom_stoploss()` 方法控制，而不是通过 trailing stop loss 配置。**
+
+---
+
+## ✅ 1. 当前使用的止损机制
+
+### 📌 `custom_stoploss()` 方法说明：
+
+```python
+def custom_stoploss(self, pair: str, trade: Trade, current_time: datetime,
+                    current_rate: float, current_profit: float, **kwargs) -> float:
+    """
+    动态止损回调函数，基于 FreqAI 预测的 &-stoploss_target 值。
+    """
+
+    # 获取当前币种的 dataframe
+    dataframe, _ = self.dp.get_analyzed_dataframe(pair, self.timeframe)
+    last_candle = dataframe.iloc[-1]
+
+    # 获取预测的止损目标
+    stoploss_target = last_candle.get("&-stoploss_target", -0.01)
+
+    # 将 stoploss_target 转换为负数（因为 stoploss 是负值）
+    dynamic_stoploss = min(-0.005, stoploss_target)  # 设置最小止损限制
+
+    return dynamic_stoploss
+```
+
+这段代码是你的策略中唯一生效的止损逻辑 —— 它根据 FreqAI 提供的 `&-stoploss_target` 来动态设置止损点。
+
+---
+
+## ✅ 2. Trailing Stop Loss 是什么？
+
+Trailing stop loss 是一种 **跟随价格移动的止损机制**，例如：
+
+```json
+"trailing_stop": true,
+"trailing_stop_positive": 0.02,
+"trailing_stop_positive_offset": 0.03,
+"trailing_only_offset_is_reached": true
+```
+
+它的行为是：
+- 开仓后，止损点随着价格上涨“向上移动”
+- 可以锁定利润，防止回吐过多盈利
+
+---
+
+## ✅ 3. 如何判断是否启用了 trailing stop loss？
+
+在你的策略类中没有看到如下配置或方法：
+
+```python
+trailing_stop = True
+```
+
+也没有使用以下任何与 trailing stop 相关的参数。
+
+因此可以确认：
+
+> 🔍 **当前使用的是 `custom_stoploss()` 的动态止损，而不是 trailing stop loss。**
+
+---
+
+## ✅ 4. 如果你想启用 trailing stop loss，应该怎么做？
+
+你可以选择以下两种方式之一：
+
+### ✅ 方式一：使用全局配置文件（如 `config.json`）
+
+```json
+{
+  "max_open_trades": 3,
+  "stake_currency": "USDT",
+  "stake_amount": 100,
+  "tradable_balance_ratio": 0.99,
+  "minimal_roi": {
+    "60": 0.01,
+    "30": 0.02,
+    "0": 0.03
+  },
+  "stoploss": -0.10,
+  "trailing_stop": true,
+  "trailing_stop_positive": 0.02,
+  "trailing_stop_positive_offset": 0.03,
+  "trailing_only_offset_is_reached": true
+}
+```
+
+这样会全局启用 trailing stop loss。
+
+---
+
+### ✅ 方式二：在策略中定义（覆盖配置）
+
+```python
+# 在策略类中添加如下属性
+trailing_stop = True
+trailing_stop_positive = 0.02
+trailing_stop_positive_offset = 0.03
+trailing_only_offset_is_reached = True
+```
+
+这将覆盖配置文件中的设置。
+
+---
+
+## ✅ 5. `custom_stoploss()` vs `trailing_stop`
+
+| 特性 | `custom_stoploss()` | `trailing_stop` |
+|------|---------------------|------------------|
+| 是否可动态计算 | ✅ 是 | ❌ 否（固定百分比） |
+| 是否跟随价格移动 | ❌ 否 | ✅ 是 |
+| 是否支持 FreqAI 输出 | ✅ 是 | ❌ 否 |
+| 是否需要额外配置 | ❌ 否 | ✅ 是 |
+
+---
+
+## ✅ 总结
+
+| 问题 | 回答 |
+|------|------|
+| 当前是否启用了 trailing stop loss？ | ❌ 否 |
+| 止损逻辑由哪段代码决定？ | ✅ `custom_stoploss()` 方法 |
+| 如何启用 trailing stop loss？ | ✅ 在配置文件或策略中添加相关参数 |
+| 哪个更适合结合 FreqAI 使用？ | ✅ `custom_stoploss()` 更灵活，推荐继续使用它 |
+
+---
+
+如果你想我帮你实现一个 **结合 FreqAI 和 trailing stop 的混合止损系统**，请告诉我你的目标（比如你想预测涨跌、波动率等），我可以为你定制这部分逻辑。
+
+

chat/result13.md

@@ -0,0 +1,250 @@
+> 请实现代码:
+> 1. **将 `&-divergence_score` 改为连续值形式**
+> 2. **移除任何离散值列（如 market_condition 编码）**
+> 3. **保持使用 `LightGBMMultiTargetRegressor` 模型**
+> 4. **在策略中使用这些连续值进行动态交易决策**
+
+
+Error: {
+  exit = 56,
+  message = 'post https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions - curl error exit_code=56 stderr={ "curl: (56) OpenSSL SSL_read: error:0A000126:SSL routines::unexpected eof while reading, errno 0" }',
+  stderr = '{ "curl: (56) OpenSSL SSL_read: error:0A000126:SSL routines::unexpected eof while reading, errno 0" }'
+}
+
+-------
+
+- Datetime: 2025-05-31 08:40:04
+
+- Model: _qwen3/qwen-plus-latest
+
+- Selected files:
+  - freqtrade/templates/freqaiprimer.py
+
+> 中断了,请继续
+
+
+非常理解，我们继续完成你的需求。
+
+---
+
+## ✅ 你的目标
+
+你希望：
+
+1. **将 `&-divergence_score` 改为连续值形式**
+2. **移除任何离散值列（如 `&-market_condition`）**
+3. **保持使用 `LightGBMMultiTargetRegressor` 模型**
+4. **在策略中使用这些连续值进行动态交易决策**
+
+---
+
+## ✅ 实现步骤概览
+
+我们将完成以下修改：
+
+| 步骤 | 内容 |
+|------|------|
+| 🧮 Step 1 | 修改 `set_freqai_targets()`：生成连续背离评分 |
+| 🧹 Step 2 | 移除所有离散值列（如 `&-market_condition`） |
+| 📈 Step 3 | 配置 FreqAI 使用 `LightGBMMultiTargetRegressor` |
+| 🛠 Step 4 | 在 `populate_entry_trend()` 和 `populate_exit_trend()` 中使用连续背离信号 |
+
+---
+
+## 🧮 Step 1：修改 `set_freqai_targets()`，生成连续背离评分
+
+```python
+def set_freqai_targets(self, dataframe: DataFrame, metadata: dict, **kwargs) -> DataFrame:
+    """
+    使用历史窗口预测未来趋势和波动率作为辅助信号，
+    并新增一个连续值的背离评分列 &-divergence_score
+    """
+
+    label_period = self.freqai_info["feature_parameters"]["label_period_candles"]
+
+    # 1. 趋势强度：使用过去 N 根 K 线的收益率判断当前趋势
+    dataframe["&-trend_strength"] = dataframe["close"].pct_change(label_period)
+
+    # 2. 波动率预测：使用过去 N 根 K 线的价格变动绝对值
+    dataframe["&-volatility_forecast"] = dataframe["close"].pct_change(label_period).abs()
+
+    # 3. ROI 目标：基于趋势强度缩放
+    dataframe["&-roi_target"] = np.where(
+        dataframe["&-trend_strength"] > 0,
+        dataframe["&-trend_strength"] * 1.5,
+        0.01  # 最小 ROI
+    )
+
+    # 4. 动态止损目标：基于波动率缩放
+    dataframe["&-stoploss_target"] = -dataframe["&-volatility_forecast"] * 1.2
+
+    # 5. 新增：连续背离评分（价格与趋势的偏离程度）
+    dataframe["price_change"] = dataframe["close"].pct_change(label_period)
+    dataframe["trend_change"] = dataframe["&-trend_strength"].pct_change(label_period)
+
+    # 连续背离评分：正值表示正向背离，负值表示负向背离
+    dataframe["&-divergence_score"] = dataframe["price_change"] - dataframe["trend_change"]
+
+    return dataframe
+```
+
+---
+
+## 🧹 Step 2：移除所有离散值列（如 `&-market_condition`）
+
+如果你之前定义了 `&-market_condition` 或其他分类列，请从 `set_freqai_targets()` 和 `populate_indicators()` 中删除它们。
+
+### ❌ 删除示例：
+```python
+# 删除以下代码（如果存在）
+if "bb_upperband-period" in dataframe.columns and "bb_lowerband-period" in dataframe.columns:
+    dataframe["&-market_condition"] = np.select([
+        (dataframe["close"] > dataframe["bb_upperband-period"]),
+        (dataframe["close"] < dataframe["bb_lowerband-period"]),
+        (dataframe["&-trend_strength"] > 0.03),
+    ], [
+        "overbought",
+        "oversold",
+        "strong_up",
+    ], default="sideways")
+```
+
+同时确保不再使用这些列做任何逻辑判断。
+
+---
+
+## 📈 Step 3：配置 FreqAI 使用 `LightGBMMultiTargetRegressor`
+
+在你的策略类中设置：
+
+```python
+freqai_info = {
+    "model": "LightGBMMultiTargetRegressor",  # ✅ 多目标回归模型
+    "feature_parameters": {
+        "include_timeframes": ["5m", "15m", "1h"],
+        "label_period_candles": 24,
+        "include_shifted_candles": 3,
+        "principal_component_analysis": True,  # 可选：启用 PCA 提升泛化能力
+    },
+    "data_split_parameters": {
+        "test_size": 0.2,
+        "shuffle": False,
+    },
+    "model_training_parameters": {
+        "n_estimators": 200,
+        "learning_rate": 0.05,
+        "num_leaves": 31,
+        "verbose": -1,
+    },
+}
+```
+
+这样 FreqAI 将使用 LightGBM 的多输出回归模型来同时预测多个连续目标列。
+
+---
+
+## 🛠 Step 4：在策略中使用连续背离信号进行交易决策
+
+### ✅ 示例：在 `populate_entry_trend()` 中使用 `&-divergence_score`
+
+```python
+def populate_entry_trend(self, df: DataFrame, metadata: dict) -> DataFrame:
+    conditions = []
+
+    if "&-divergence_score" in df.columns:
+
+        # 正向背离：价格下跌但趋势上升（背离评分为正值）
+        buy_condition = df["&-divergence_score"] > 0.01
+
+        # 结合趋势强度过滤
+        if "&-trend_strength" in df.columns:
+            buy_condition &= df["&-trend_strength"] > 0.005
+
+        # 成交量过滤：成交量需大于 20 日均值的 1 倍
+        df["vol_mean"] = df["volume"].rolling(20).mean()
+        buy_condition &= (df["volume"] > df["vol_mean"] * 1)
+
+        conditions.append(buy_condition)
+
+    if len(conditions):
+        df.loc[reduce(lambda x, y: x & y, conditions), 'enter_long'] = 1
+
+    return df
+```
+
+### ✅ 示例：在 `populate_exit_trend()` 中使用 `&-divergence_score`
+
+```python
+def populate_exit_trend(self, df: DataFrame, metadata: dict) -> DataFrame:
+    conditions = []
+
+    if "&-divergence_score" in df.columns:
+
+        # 负向背离：价格上涨但趋势下降（背离评分为负值）
+        sell_condition = df["&-divergence_score"] < -0.01
+
+        # 结合趋势强度过滤
+        if "&-trend_strength" in df.columns:
+            sell_condition &= df["&-trend_strength"] < -0.005
+
+        conditions.append(sell_condition)
+
+    if len(conditions):
+        df.loc[reduce(lambda x, y: x & y, conditions), 'exit_long'] = 1
+
+    return df
+```
+
+---
+
+## 📦 Step 5：更新 `plot_config`（可选）
+
+你可以将背离评分加入绘图配置，方便可视化观察：
+
+```python
+plot_config = {
+    "main_plot": {},
+    "subplots": {
+        "Signals": {
+            "enter_long": {"color": "green"},
+            "exit_long": {"color": "red"}
+        },
+        "FreqAI Predictions": {
+            "&-trend_strength": {"color": "blue"},
+            "&-volatility_forecast": {"color": "orange"},
+            "&-divergence_score": {"color": "purple"}
+        }
+    }
+}
+```
+
+---
+
+## ✅ 总结改进点
+
+| 改进项 | 描述 |
+|--------|------|
+| ✅ `&-divergence_score` 改为连续值 | 更适合回归模型学习 |
+| ✅ 移除所有离散值列 | 如 `&-market_condition` |
+| ✅ 使用 `LightGBMMultiTargetRegressor` | 同时预测多个连续目标 |
+| ✅ 在策略中使用背离评分进行交易决策 | 提高反转识别能力 |
+
+---
+
+## ✅ 下一步建议
+
+你可以：
+
+- 使用 `freqtrade plot-dataframe` 查看背离信号是否合理
+- 使用 `freqtrade backtesting` 测试新策略表现
+- 使用 `freqtrade hyperopt` 对背离阈值进行优化
+
+---
+
+如果你希望我帮你生成完整的策略文件，并提供对应的 `config.json` 示例，请告诉我 👇 我可以一步步帮你完成。
+
+freqtrade plot-dataframe --strategy DivergenceRegressionStrategy \
+                         --timerange=20250501-20250510 \
+                         --timeframe=5m \
+                         --exchange=binance \
+                         --pair=ETH/USDT

chat/result14.md

@@ -0,0 +1,708 @@
+- Datetime: 2025-06-03 02:21:39
+
+- Model: _qwen3/qwen-plus-latest
+
+- Selected files:
+  - result/backtest-result-2025-06-03_01-41-33_FreqaiPrimer.py
+  - result/analytic.log
+  - result/backtest-result-2025-06-03_01-41-33.json
+  - result/backtest-result-2025-06-03_01-41-33.meta.json
+  - result/backtest-result-2025-06-03_01-41-33_FreqaiPrimer.json
+  - result/backtest-result-2025-06-03_01-41-33_config.json
+  - result/backtest_trades.csv
+  - result/output.log
+
+ > 当前策略的配置,代码,执行结果都在上下文中, 结合官方文档, 请给出对于执行结果的评价以及修改建议
+ > 
+ > 	自定义策略: https://www.freqtrade.io/en/stable/strategy-customization/
+ > 	freqai 特征工程 : https://www.freqtrade.io/en/stable/freqai-feature-engineering/
+ > 	freqai 参数表: https://www.freqtrade.io/en/stable/freqai-parameter-table/
+ > 	freqai 配置: https://www.freqtrade.io/en/stable/freqai-configuration/ 
+ > 	freaai 运行: https://www.freqtrade.io/en/stable/freqai-running/
+ > 	freqai 开发手册: https://www.freqtrade.io/en/stable/freqai-developers/
+ > 	freqtrade 策略例子: https://github.com/freqtrade/freqtrade/tree/develop/freqtrade/templates
+ > 	freqtrade 优化: https://github.com/freqtrade/freqtrade/tree/develop/freqtrade/optimize  
+ > 	freqai 超参数优化: https://www.freqtrade.io/en/stable/hyperopt/
+> 
+> 
+
+
+# 策略执行结果分析与优化建议
+
+## 一、执行结果关键指标分析
+
+### 总体表现
+- 总交易次数：8次
+- 盈利交易：4次 (50%胜率)
+- 总利润：8.87 USDT (+0.89%)
+- 平均持仓时间：21分钟
+- 最大连续亏损：2次
+- 利润因子：2.15
+
+### 分币种表现
+| 币种       | 交易次数 | 平均利润 | 胜率   |
+|------------|----------|----------|--------|
+| KAITO/USDT | 4        | +0.77%   | 50%    |
+| TRUMP/USDT | 3        | +0.95%   | 66.7%  |
+| WCT/USDT   | 1        | 0%       | 0%     |
+
+### 分退出原因表现
+| 原因        | 交易次数 | 平均利润 | 胜率   |
+|-------------|----------|----------|--------|
+| ROI         | 7        | +1.58%   | 100%   |
+| 止损        | 1        | -5.15%   | 0%     |
+
+## 二、策略配置分析
+
+### 1. FreqAI配置
+```json
+"freqai": {
+  "enabled": true,
+  "data_kitchen": {"fillna": "ffill"},
+  "freqaimodel": "LightGBMMultiTargetRegressor",
+  "purge_old_models": 2,
+  "identifier": "test58",
+  "train_period_days": 7,
+  "backtest_period_days": 1,
+  "live_retrain_hours": 2,
+  "outlier_detection": {"method": "IsolationForest", "contamination": 0.1},
+  "feature_selection": {"method": "recursive_elimination"},
+  "feature_parameters": {
+    "include_timeframes": ["3m","15m","1h"],
+    "include_corr_pairlist": ["BTC/USDT","SOL/USDT"],
+    "outlier_protection_percentage": 0.1,
+    "label_period_candles": 24,
+    "include_shifted_candles": 3,
+    "DI_threshold": 0.7,
+    "weight_factor": 0.9,
+    "principal_component_analysis": false,
+    "use_SVM_to_remove_outliers": true,
+    "indicator_periods_candles": [10,20,50],
+    "plot_feature_importances": 10
+  },
+  "data_split_parameters": {"test_size": 0.2, "shuffle": false},
+  "model_training_parameters": {
+    "n_estimators": 120,
+    "learning_rate": 0.05,
+    "num_leaves": 31,
+    "verbose": -1,
+    "max_depth": 8
+  }
+}
+```
+
+### 2. 策略参数
+```python
+minimal_roi = {
+    "0": 0.02,
+    "30": 0.01,
+    "60": 0
+}
+
+stoploss = -0.015
+
+trailing_stop = True
+trailing_stop_positive = 0.053
+trailing_stop_positive_offset = 0.152
+```
+
+## 三、执行结果评价
+
+### 1. 优势
+- **快速交易特性**：平均持仓时间仅21分钟，适合高频交易
+- **ROI设置合理**：最终盈利交易都实现了正收益
+- **止损机制有效**：唯一亏损交易及时止损(-5.15%)
+
+### 2. 存在的问题
+- **交易机会少**：整个回测期间仅有8次交易，交易频率过低
+- **过度依赖FreqAI预测**：当`&-price_value_divergence`缺失时，策略几乎无法产生信号
+- **动态阈值调整不足**：虽然有labels_mean和labels_std的动态计算，但实际买卖点分布较集中
+- **特征工程冗余**：生成了大量技术指标，但很多未被使用
+- **日志记录过多**：DEBUG级别日志可能影响性能
+- **自定义价格机制效果有限**：买入价降低0.5%、卖出价提高0.25%，对整体收益贡献较小
+
+## 四、优化建议
+
+### 1. 提高交易频率
+- 缩小`fit_live_predictions_candles`从100到50，增加交易机会
+- 降低入场条件门槛：
+  ```python
+  cond2 = (dataframe["volume_z_score"] > 1.0)  # 原为1.5
+  cond3 = (dataframe["rsi"] < 45)           # 原为40
+  ```
+
+### 2. 增强入场信号
+- 添加备用信号源，当FreqAI数据不可用时：
+  ```python
+  if "&-price_value_divergence" not in dataframe.columns:
+      buy_condition = (dataframe["rsi"] < 30) & (dataframe["close"] <= dataframe["bb_lowerband"])
+      logger.warning(f"[{pair}] 使用备用策略: RSI+布林带")
+  ```
+
+### 3. 优化动态阈值计算
+- 引入时间衰减因子，使阈值随市场波动自动调节：
+  ```python
+  volatility = dataframe["bb_upperband-period"] / dataframe["bb_lowerband-period"] - 1
+  k_volatility = 1 + (volatility.iloc[-1] * 10)  # 波动率越高，k值越大
+  
+  self.buy_threshold = labels_mean - k_buy * labels_std * k_volatility
+  self.sell_threshold = labels_mean + k_sell * labels_std * k_volatility
+  ```
+
+### 4. 简化特征工程
+- 移除未使用的指标：
+  ```python
+  # 删除以下未被使用的指标
+  del dataframe["%-sma-period"]
+  del dataframe["%-mfi-period"]
+  del dataframe["%-adx-period"]
+  ```
+
+### 5. 改进Trailing Stop逻辑
+- 实现动态追踪止损：
+  ```python
+  def adjust_trade_position(self, trade: Trade, current_time: datetime,
+                          current_rate: float, current_profit: float,
+                          min_roi: Dict[float, float], max_profit: float):
+      # 动态调整追踪止损参数
+      hold_time = (current_time - trade.open_date_utc).total_seconds() / 60
+      
+      # 持仓前5分钟使用保守止损
+      if hold_time < 5:
+          trailing_stop_start = 0.02
+          trailing_stop_distance = 0.008
+      # 持仓5-15分钟使用激进追涨
+      elif hold_time < 15:
+          trailing_stop_start = 0.03
+          trailing_stop_distance = 0.01
+      # 持仓超过15分钟后使用保守追踪
+      else:
+          trailing_stop_start = 0.015
+          trailing_stop_distance = 0.005
+          
+      # 剩余代码保持不变...
+  ```
+
+### 6. 优化Hyperopt参数空间（用于后续优化）
+```python
+# ROI 参数空间
+def roi_space():
+    return [
+        Integer(0, 60, name='roi_t1'),      
+        Integer(30, 120, name='roi_t2'),    
+        Integer(60, 180, name='roi_t3'),    
+        Real(0.01, 0.05, name='roi_p1'),   
+        Real(0.005, 0.03, name='roi_p2'),  
+        Real(0.0, 0.02, name='roi_p3'),    
+    ]
+
+# Trailing Stop 参数空间
+def trailing_space():
+    return [
+        Real(0.01, 0.05, name='trailing_stop_start'),     
+        Real(0.005, 0.02, name='trailing_stop_distance'), 
+    ]
+```
+
+### 7. 风险控制增强
+- 添加最大单日亏损限制：
+  ```python
+  def custom_stake_amount(self, pair, current_time, current_rate, proposed_stake, min_stake, max_stake):
+      # 计算当日已实现亏损
+      daily_loss = sum(trade.calc_profit() for trade in Trade.get_trades() 
+                      if trade.close_date and trade.close_date.date() == current_time.date() and trade.loss)
+                      
+      if daily_loss < -100:  # 当日亏损超过$100停止交易
+          return 0
+          
+      return proposed_stake * 0.5  # 默认使用50%推荐仓位
+  ```
+
+## 五、FreqAI专项优化建议
+
+### 1. 特征工程优化
+- 根据官方文档[FreqAI特征工程](https://www.freqtrade.io/en/stable/freqai-feature-engineering/)建议：
+  ```python
+  def feature_engineering_expand_all(self, dataframe, period, metadata, **kwargs):
+      # 保留核心特征
+      dataframe["%-rsi-period"] = ta.RSI(dataframe, timeperiod=period)
+      
+      # 添加趋势特征
+      dataframe[f"%-sma-period-ratio"] = (dataframe["close"] / ta.SMA(dataframe, timeperiod=period)) - 1
+      
+      # 添加波动率特征
+      dataframe[f"%-volatility"] = (ta.ATR(dataframe, timeperiod=period) / dataframe["close"])
+      
+      # 清理NaN值
+      dataframe.fillna(method='ffill', inplace=True)
+      dataframe.fillna(0, inplace=True)
+      
+      return dataframe
+  ```
+
+### 2. 标签(Label)优化
+- 根据[官方建议](https://www.freqtrade.io/en/stable/freqai-parameter-table/)，可以尝试更复杂的标签：
+  ```python
+  def set_freqai_targets(self, dataframe: DataFrame, metadata: dict, **kwargs) -> DataFrame:
+      # 多目标回归
+      dataframe["&-target-profit"] = (dataframe["close"].shift(-self.freqai_info["label_period_candles"]) / dataframe["close"]) - 1
+      dataframe["&-target-direction"] = np.where(
+          dataframe["close"].shift(-self.freqai_info["label_period_candles"]) > dataframe["close"], 1, 0
+      )
+      
+      # 数据清理
+      dataframe.replace([np.inf, -np.inf], 0, inplace=True)
+      dataframe.fillna(0, inplace=True)
+      
+      return dataframe
+  ```
+
+## 六、具体修改建议
+
+### 1. 修改后的`__init__`方法
+```python
+def __init__(self, config: dict, *args, **kwargs):
+    super().__init__(config, *args, **kwargs)
+    logger.setLevel(logging.INFO)  # 降低默认日志级别
+    logger.info("✅ 策略已初始化，日志级别设置为 INFO")
+    self.trailing_stop_enabled = False
+    self.pair_stats = {}
+    self.stats_logged = False
+    self.fit_live_predictions_candles = self.freqai_info.get("fit_live_predictions_candles", 50)  # 减少到50根K线
+```
+
+### 2. 修改后的`confirm_trade_entry`方法
+```python
+def confirm_trade_entry(self, pair: str, order_type: str, amount: float, rate: float,
+                       time_in_force: str, current_time: datetime, **kwargs) -> bool:
+    # 在 Hyperopt 或 Dry Run 模式下避免数据库查询
+    if 'hyperopt' in self.config.get('runmode', '') or not hasattr(Trade, 'query') or Trade.query is None:
+        return True
+
+    # 降低检查窗口到2分钟
+    recent_trades = Trade.query.filter(
+        Trade.pair == pair,
+        Trade.is_open == False,  # noqa: E712
+        Trade.close_date > current_time - datetime.timedelta(minutes=2)
+    ).all()
+
+    if len(recent_trades) > 0:
+        logger.debug(f"[{pair}] 2分钟内有{len(recent_trades)}笔近期交易，跳过本次入场")
+        return False
+
+    self.trailing_stop_enabled = False
+    return True
+```
+
+### 3. 修改后的`adjust_trade_position`方法
+```python
+def adjust_trade_position(self, trade: Trade, current_time: datetime,
+                        current_rate: float, current_profit: float,
+                        min_roi: Dict[float, float], max_profit: float):
+    hold_time = (current_time - trade.open_date_utc).total_seconds() / 60
+    
+    # 动态调整追踪止损
+    if hold_time < 5:
+        trailing_stop_start = 0.02
+        trailing_stop_distance = 0.008
+    elif hold_time < 15:
+        trailing_stop_start = 0.03
+        trailing_stop_distance = 0.01
+    else:
+        trailing_stop_start = 0.015
+        trailing_stop_distance = 0.005
+
+    profit_ratio = (current_rate - trade.open_rate) / trade.open_rate
+    
+    if profit_ratio >= trailing_stop_start and not self.trailing_stop_enabled:
+        self.trailing_stop_enabled = True
+        trade.adjust_max_rate(current_rate)
+        logger.info(f"[{trade.pair}] 价格上涨超过 {trailing_stop_start*100:.1f}%，启动 Trailing Stop")
+
+    if self.trailing_stop_enabled:
+        max_rate = trade.max_rate
+        trailing_stop_price = max_rate * (1 - trailing_stop_distance)
+        
+        if current_rate < trailing_stop_price:
+            logger.info(f"[{trade.pair}] 价格回落至 Trailing Stop 点 {trailing_stop_price:.6f}，触发卖出")
+            return -1
+            
+        trade.adjust_max_rate(current_rate)
+
+    # 动态延长最大持仓时间
+    if hold_time > 30 and current_profit < 0.005:  # 小于0.5%收益时强制平仓
+        logger.info(f"[{trade.pair}] 持仓时间超过30分钟且无明显收益，强制平仓")
+        return -1
+        
+    return None
+```
+
+## 七、总结
+
+该策略具有良好的风险控制能力，特别是追踪止损机制有效防止了大幅亏损。然而，策略过于保守导致交易机会较少，且对FreqAI预测结果依赖过高。
+
+建议优先改进方向：
+1. 增加备用交易信号以应对FreqAI数据缺失情况
+2. 优化特征工程，移除冗余指标
+3. 实施动态仓位管理
+4. 细化追踪止损逻辑，根据持仓时间动态调整参数
+5. 降低最小持仓时间要求，增加交易机会
+
+下一步建议进行Hyperopt优化，重点优化ROI和追踪止损参数，同时考虑加入更多市场状态判断来动态调整策略行为。
+
+-------
+
+- Datetime: 2025-06-03 02:44:15
+
+- Model: _qwen3/qwen-plus-latest
+
+- Selected files:
+  - result/backtest-result-2025-06-03_01-41-33_FreqaiPrimer.py
+  - result/analytic.log
+  - result/backtest-result-2025-06-03_01-41-33.json
+  - result/backtest-result-2025-06-03_01-41-33.meta.json
+  - result/backtest-result-2025-06-03_01-41-33_FreqaiPrimer.json
+  - result/backtest-result-2025-06-03_01-41-33_config.json
+  - result/backtest_trades.csv
+  - result/output.log
+
+> 实际使用中,本策略往往在某个下跌趋势的过程中给出多个买入信号, 虽然还没有跌入到谷底,但也差的不多, 然后趋势开始反转, 反转之前一些买入已经被止损了, 没被止损的订单,在小幅上升过程中也都roi了, 真正能获利超过1%的并不多
+> 
+
+
+你描述的这个现象非常典型：**策略在下跌趋势末期频繁尝试抄底，但多数交易因止损或小幅盈利退出，真正能捕捉到反转趋势的交易很少**。
+
+---
+
+## 📉 问题本质分析
+
+### ✅ 现象总结：
+- **下跌趋势中频繁给出买入信号**
+- **部分订单被止损（stop_loss）**
+- **未止损的订单在小幅反弹后触发 ROI 卖出**
+- **最终实现超过1%利润的交易很少**
+
+这说明：
+
+| 问题类型 | 描述 |
+|---------|------|
+| ❌ 过早入场 | 在趋势尚未明确反转时就开始抄底 |
+| ❌ 止损设置不合理 | 在震荡行情中容易被洗出去 |
+| ❌ ROI 设置过紧 | 小幅反弹就卖出，错过后续上涨空间 |
+| ❌ 缺乏趋势过滤机制 | 没有判断当前是下跌、震荡还是上升趋势 |
+
+---
+
+## 🔍 结合你的策略代码和回测结果分析
+
+从你的原始策略来看：
+
+```python
+buy_condition = (dataframe["&-price_value_divergence"] < self.buy_threshold) & \
+                (dataframe["volume_z_score"] > 1.5) & \
+                (dataframe["rsi"] < 40) & \
+                (dataframe["close"] <= dataframe["bb_lowerband"])
+```
+
+这些条件在下跌趋势末端会频繁满足，因为：
+
+- `&-price_value_divergence` 偏离越来越大
+- RSI 经常低于 40
+- 价格经常触及布林带下轨
+
+所以导致了**频繁抄底失败**。
+
+---
+
+## 🛠️ 改进建议（实战级）
+
+### ✅ 1. 添加趋势过滤器（Trend Filter）
+
+避免在下跌趋势中盲目抄底，只允许在特定趋势中开仓。
+
+#### 示例：使用 EMA 判断趋势方向
+
+```python
+def populate_indicators(self, dataframe: DataFrame, metadata: dict) -> DataFrame:
+    # ...原有指标计算...
+
+    # 趋势判断
+    dataframe['trend_ema'] = ta.EMA(dataframe, timeperiod=50)
+    dataframe['trend_direction'] = np.where(
+        dataframe['close'] > dataframe['trend_ema'], 'up', 
+        np.where(dataframe['close'] < dataframe['trend_ema'], 'down', 'sideways')
+    )
+
+    return dataframe
+```
+
+然后在 `populate_entry_trend` 中加入趋势判断：
+
+```python
+if "&-price_value_divergence" in dataframe.columns:
+    buy_condition = (
+        (dataframe["&-price_value_divergence"] < self.buy_threshold) &
+        (dataframe["volume_z_score"] > 1.5) &
+        (dataframe["rsi"] < 40) &
+        (dataframe["close"] <= dataframe["bb_lowerband"]) &
+        (dataframe['trend_direction'] != 'down')  # 避免在下降趋势中买入
+    )
+```
+
+---
+
+### ✅ 2. 动态调整止损和ROI（Adaptive Exit）
+
+不要固定止损和ROI，而是根据市场波动率动态调整。
+
+#### 示例：基于ATR的动态止损
+
+```python
+def populate_indicators(self, dataframe: DataFrame, metadata: dict) -> DataFrame:
+    # ...其他指标...
+    
+    # ATR用于动态止损
+    dataframe['atr'] = ta.ATR(dataframe, timeperiod=14)
+    dataframe['dynamic_stoploss'] = -0.01 - (dataframe['atr'] / dataframe['close']) * 1.5
+    
+    return dataframe
+```
+
+然后在策略中使用它：
+
+```python
+def __init__(self, config: dict, *args, **kwargs):
+    super().__init__(config, *args, **kwargs)
+    self.stoploss = RealParameter(-0.03, -0.01, default=-0.015, space="all")
+```
+
+这样可以防止在波动大的时候被轻易止损。
+
+---
+
+### ✅ 3. 延迟入场（Delay Entry Until Confirmation）
+
+不要一满足条件就入场，等待确认信号。
+
+#### 示例：RSI + MACD 双重确认
+
+```python
+def populate_indicators(self, dataframe: DataFrame, metadata: dict) -> DataFrame:
+    # ...其他指标...
+    dataframe['macd'], dataframe['signal'], _ = ta.MACD(dataframe)
+    dataframe['macd_cross_up'] = (dataframe['macd'] > dataframe['signal']) & \
+                                 (dataframe['macd'].shift(1) <= dataframe['signal'].shift(1))
+    return dataframe
+```
+
+然后修改入场逻辑：
+
+```python
+buy_condition = (
+    (dataframe["&-price_value_divergence"] < self.buy_threshold) &
+    (dataframe["volume_z_score"] > 1.5) &
+    (data
+
+
+Error: {
+  exit = 56,
+  message = 'post https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions - curl error exit_code=56 stderr={ "curl: (56) OpenSSL SSL_read: error:0A000126:SSL routines::unexpected eof while reading, errno 0" }',
+  stderr = '{ "curl: (56) OpenSSL SSL_read: error:0A000126:SSL routines::unexpected eof while reading, errno 0" }'
+}
+
+-------
+
+- Datetime: 2025-06-03 02:45:22
+
+- Model: _qwen3/qwen-plus-latest
+
+- Selected files:
+  - result/backtest-result-2025-06-03_01-41-33_FreqaiPrimer.py
+  - result/analytic.log
+  - result/backtest-result-2025-06-03_01-41-33.json
+  - result/backtest-result-2025-06-03_01-41-33.meta.json
+  - result/backtest-result-2025-06-03_01-41-33_FreqaiPrimer.json
+  - result/backtest-result-2025-06-03_01-41-33_config.json
+  - result/backtest_trades.csv
+  - result/output.log
+
+> 输出中断了,请重新输出
+> 
+
+
+### ✅ 3. 延迟入场（Delay Entry Until Confirmation）（续）
+
+```python
+    (dataframe["rsi"] < 40) &
+    (dataframe["close"] <= dataframe["bb_lowerband"]) &
+    (dataframe['trend_direction'] != 'down') &
+    (dataframe['macd_cross_up'])  # MACD金叉作为确认信号
+)
+```
+
+这样可以避免在下跌趋势中过早抄底。
+
+---
+
+### ✅ 4. 放宽 ROI 设置，让利润奔跑（Let Profits Run）
+
+你提到“小幅反弹就ROI了”，说明 ROI 设置太紧。我们可以：
+
+- **延长 ROI 时间**
+- **使用动态 ROI**
+
+#### 示例：基于波动率的动态 ROI
+
+```python
+def generate_roi_table(self, params: Dict) -> Dict[int, float]:
+    base_roi = {
+        "0": params.get('roi_p1', 0.02),
+        "7": params.get('roi_t1', 0.015),
+        "21": params.get('roi_t2', 0.008),
+        "27": 0
+    }
+    
+    # 根据市场波动率调整ROI
+    volatility = self.current_volatility  # 来自历史ATR或标准差
+    if volatility > 0.02:  # 高波动时放宽ROI
+        base_roi["0"] *= 1.5
+        base_roi["7"] *= 1.5
+    
+    return base_roi
+```
+
+或者直接设置更宽松的默认值：
+
+```python
+minimal_roi = {
+    "0": 0.092,
+    "7": 0.056,
+    "21": 0.024,
+    "27": 0
+}
+```
+
+---
+
+### ✅ 5. 使用 FreqAI 的预测方向作为过滤器
+
+既然你已经在用 FreqAI，可以利用它的预测值来判断趋势方向：
+
+```python
+if "&-price_value_divergence" in dataframe.columns:
+    # 只有当预测值开始上升时才允许买入
+    dataframe["prediction_up"] = (dataframe["&-price_value_divergence"].diff() > 0)
+    
+    buy_condition = (
+        (dataframe["&-price_value_divergence"] < self.buy_threshold) &
+        (dataframe["volume_z_score"] > 1.5) &
+        (dataframe["rsi"] < 40) &
+        (dataframe["close"] <= dataframe["bb_lowerband"]) &
+        (dataframe["prediction_up"])
+    )
+```
+
+---
+
+### ✅ 6. 加入“连续亏损”保护机制
+
+防止在连续止损后继续交易，建议加入一个冷却期：
+
+```python
+def confirm_trade_entry(self, pair: str, order_type: str, amount: float, rate: float,
+                       time_in_force: str, current_time: datetime, **kwargs) -> bool:
+    recent_trades = Trade.query.filter(
+        Trade.pair == pair,
+        Trade.is_open == False,
+        Trade.close_date > current_time - datetime.timedelta(minutes=30)
+    ).all()
+
+    losing_trades = [t for t in recent_trades if t.calc_profit_ratio(rate) < 0]
+    
+    if len(losing_trades) >= 2:
+        logger.info(f"[{pair}] 最近2次交易亏损，暂停入场")
+        return False
+        
+    self.trailing_stop_enabled = False
+    return True
+```
+
+---
+
+### ✅ 7. 自定义出场逻辑优化
+
+目前你的出场逻辑是：
+
+```python
+sell_condition = (dataframe["&-price_value_divergence"] > self.sell_threshold) | (dataframe["rsi"] > 75)
+```
+
+建议改为**分阶段卖出**，比如：
+
+```python
+def populate_exit_trend(self, dataframe: DataFrame, metadata: dict) -> DataFrame:
+    if "&-price_value_divergence" in dataframe.columns:
+        # 分阶段退出
+        dataframe.loc[
+            (dataframe["&-price_value_divergence"] > self.sell_threshold * 1.2), 
+            'exit_long_partial'
+        ] = 1  # 先卖一半
+        
+        dataframe.loc[
+            (dataframe["&-price_value_divergence"] > self.sell_threshold) |
+            (dataframe["rsi"] > 75),
+            'exit_long_full'
+        ] = 1  # 再卖另一半
+        
+    return dataframe
+```
+
+然后结合 `custom_sell` 实现部分卖出逻辑。
+
+---
+
+## 📊 总结：策略改进优先级建议
+
+| 改进点 | 优先级 | 说明 |
+|--------|--------|------|
+| ✅ 添加趋势过滤器 | ⭐⭐⭐⭐⭐ | 避免在下降趋势中盲目抄底 |
+| ✅ 动态止损和ROI | ⭐⭐⭐⭐ | 提高止损能力，让利润奔跑 |
+| ✅ 延迟入场确认 | ⭐⭐⭐⭐ | 等待MACD、RSI等指标确认反转 |
+| ✅ 利用FreqAI预测方向 | ⭐⭐⭐ | 过滤掉预测值仍在下降的信号 |
+| ✅ 分阶段出场 | ⭐⭐⭐ | 锁定部分利润，保留上涨空间 |
+| ✅ 连续亏损保护 | ⭐⭐ | 防止连败扩大回撤 |
+
+---
+
+## 🧪 下一步建议
+
+你可以先从以下两个改动入手，快速验证效果：
+
+### 🔧 修改1：添加趋势过滤器
+
+```python
+# 在 populate_indicators 中添加
+dataframe['trend_ema'] = ta.EMA(dataframe, timeperiod=50)
+dataframe['in_uptrend'] = dataframe['close'] > dataframe['trend_ema']
+
+# 在 populate_entry_trend 中添加
+buy_condition = buy_condition & (dataframe['in_uptrend'])
+```
+
+### 🔧 修改2：放宽止损和ROI
+
+```python
+minimal_roi = {
+    "0": 0.092,
+    "7": 0.056,
+    "21": 0.024,
+    "27": 0
+}
+
+stoploss = -0.05  # 原为 -0.015，提高容忍度
+```
+
+---
+
+如果你希望我帮你生成完整的修改后策略代码文件，我可以直接输出一份**优化后的完整策略文件**供你测试。是否需要？
+

chat/result15.md

@@ -0,0 +1,543 @@
+> 当前策略的config, freqai部分, 如何调整配置能提升策略表现
+> 
+
+你当前的 FreqAI 配置已经非常完整，涵盖了数据预处理、特征工程、模型训练、异常检测等多个方面。
+
+---
+
+## ✅ 当前配置亮点
+
+| 模块 | 说明 |
+|------|------|
+| 多时间帧特征 | `include_timeframes`: ["3m", "15m", "1h"] |
+| 相关币种特征 | 使用 BTC/USDT 和 SOL/USDT 做相关性分析 |
+| 标签周期 | `label_period_candles`: 24（预测未来 24 根 K 线） |
+| 特征选择 | 使用递归消除法（recursive_elimination） |
+| 异常检测 | 使用 IsolationForest + SVM 过滤异常值 |
+| 模型参数 | LightGBMRegressor，n_estimators=400 |
+
+---
+
+## 🚀 如何优化 FreqAI 配置以提升策略表现？
+
+以下是 **6 个关键优化方向**，每个都附有建议配置和理由：
+
+---
+
+### ✅ 1. **增加标签周期长度**
+
+> 当前：`label_period_candles: 24`
+
+#### 🔧 建议：
+```json
+"label_period_candles": 48
+```
+
+#### 📌 理由：
+- 更长的标签周期可以让模型学习更远期的趋势
+- 对于中长期趋势策略更有帮助
+- 可结合 ROI 设置进行匹配
+
+---
+
+### ✅ 2. **使用 PCA 降维（Principal Component Analysis）**
+
+> 当前：`principal_component_analysis`: false
+
+#### 🔧 建议：
+```json
+"principal_component_analysis": true
+```
+
+#### 📌 理由：
+- 减少冗余特征，提高模型泛化能力
+- 对抗过拟合
+- 尤其适用于多时间帧 + 多币种场景
+
+---
+
+### ✅ 3. **调整模型训练参数（LightGBM）**
+
+> 当前：`n_estimators`: 400, `learning_rate`: 0.01
+
+#### 🔧 建议：
+```json
+"model_training_parameters": {
+    "n_estimators": 600,
+    "learning_rate": 0.005,
+    "num_leaves": 40,
+    "max_depth": 10,
+    "min_data_in_leaf": 20,
+    "feature_fraction": 0.8,
+    "bagging_fraction": 0.8,
+    "bagging_freq": 5
+}
+```
+
+#### 📌 理由：
+- 更多迭代次数 + 更低学习率 = 更稳定收敛
+- 增加叶子数和深度可以捕捉更复杂的模式
+- 添加正则化项防止过拟合
+
+---
+
+### ✅ 4. **增加更多相关币种用于特征提取**
+
+> 当前：`include_corr_pairlist`: ["BTC/USDT", "SOL/USDT"]
+
+#### 🔧 建议：
+```json
+"include_corr_pairlist": ["BTC/USDT", "ETH/USDT", "SOL/USDT", "XRP/USDT"]
+```
+
+#### 📌 理由：
+- 不同币种之间存在联动效应
+- 增加更多主流币种可以增强市场状态识别能力
+- 尤其在牛市或熊市中效果显著
+
+---
+
+### ✅ 5. **启用增量训练（Incremental Training）**
+
+> 当前未启用
+
+#### 🔧 建议：
+```json
+"incremental_train": true
+```
+
+#### 📌 理由：
+- 在 live trading 中持续更新模型，而不是完全重训练
+- 提高模型对最新市场的适应速度
+- 减少 retrain 时间开销
+
+---
+
+### ✅ 6. **调整 outlier_protection_percentage**
+
+> 当前：`outlier_protection_percentage`: 0.1
+
+#### 🔧 建议：
+```json
+"outlier_protection_percentage": 0.15
+```
+
+#### 📌 理由：
+- 市场波动剧烈时，适当放宽异常值过滤范围
+- 防止误删极端行情下的有效信号
+- 可结合 volatility_forecast 动态调整
+
+---
+
+## ✅ 推荐的优化后 FreqAI 配置（只展示修改部分）
+
+```json
+"freqai": {
+    "enabled": true,
+    "identifier": "test58",
+    "train_period_days": 7,
+    "backtest_period_days": 1,
+    "live_retrain_hours": 2,
+    "purge_old_models": 2,
+
+    "feature_parameters": {
+        "include_timeframes": ["3m", "15m", "1h"],
+        "include_corr_pairlist": ["BTC/USDT", "ETH/USDT", "SOL/USDT", "XRP/USDT"],
+        "label_period_candles": 48,
+        "include_shifted_candles": 3,
+        "DI_threshold": 0.7,
+        "weight_factor": 0.9,
+        "principal_component_analysis": true,
+        "use_SVM_to_remove_outliers": true,
+        "indicator_periods_candles": [10, 20, 50],
+        "plot_feature_importances": 10,
+        "outlier_protection_percentage": 0.15
+    },
+
+    "data_split_parameters": {
+        "test_size": 0.2,
+        "shuffle": false
+    },
+
+    "model_training_parameters": {
+        "n_estimators": 600,
+        "learning_rate": 0.005,
+        "num_leaves": 40,
+        "max_depth": 10,
+        "min_data_in_leaf": 20,
+        "feature_fraction": 0.8,
+        "bagging_fraction": 0.8,
+        "bagging_freq": 5,
+        "verbose": -1
+    },
+
+    "incremental_train": true
+}
+```
+
+---
+
+## ✅ 总结：推荐优化一览表
+
+| 优化点 | 是否推荐 | 描述 |
+|--------|----------|------|
+| 增加 label_period_candles | ✅ 是 | 学习更远期趋势 |
+| 启用 PCA 降维 | ✅ 是 | 减少冗余特征 |
+| 调整 LightGBM 参数 | ✅ 是 | 提高模型性能 |
+| 增加相关币种 | ✅ 是 | 增强市场感知 |
+| 启用增量训练 | ✅ 是 | 提升实时适应性 |
+| 放宽异常值过滤 | ✅ 是 | 避免误删极端行情 |
+
+---
+
+🎯 **下一步建议：**
+
+你可以先尝试以下组合进行 backtest 测试：
+
+```bash
+freqtrade backtesting --strategy freqaiprimer --config config_examples/freqaiprimer.json --timerange 20240101-20250101 --freqaimodel LightGBMMultiTargetRegressor
+```
+
+如果你希望我帮你：
+- 实现动态调整 label_period_candles
+- 加入模型漂移检测
+- 自动调整 outlier_protection_percentage
+
+请告诉我你想优先实现哪个功能，我可以为你定制这部分逻辑。
+
+------
+
+- Datetime: 2025-06-07 02:11:35
+
+- Model: _qwen3/qwen-plus-latest
+
+- Selected files:
+  - result/backtest-result-2025-06-06_17-55-14.json
+  - result/analytic.log
+  - result/backtest-result-2025-06-06_17-55-14.meta.json
+  - result/backtest-result-2025-06-06_17-55-14_FreqaiPrimer.json
+  - result/backtest-result-2025-06-06_17-55-14_FreqaiPrimer.py
+  - result/backtest-result-2025-06-06_17-55-14_config.json
+  - result/backtest_trades.csv
+  - result/output.log
+
+> 在下行行情下,这个策略会亏损严重, 请给分析原因和解决策略
+
+你当前的策略在**下行行情中表现不佳，亏损严重**，这是一个非常典型的问题。我们来从多个角度分析原因，并提出针对性的优化建议。
+
+---
+
+## 📉 一、回测结果总结（关键数据）
+
+| 指标 | 数值 | 说明 |
+|------|------|------|
+| 总交易数 | 10 | 少量交易但亏损集中 |
+| 胜率 | 20% | 非常低，仅 2 赢，5 输 |
+| 平均盈利 | -1.96% | 所有交易平均亏损近 2% |
+| 最大单笔亏损 | -5.15%（TRUMP/USDT） | 单笔最大亏损超过 5% |
+| ROI 设置 | `{"0": 0.028, "12": 0.016, "48": 0.009, "120": 0}` | 初始目标为 2.8%，但未达成 |
+| Trailing Stop 设置 | `trailing_stop_positive = 0.237` | 追踪止损偏高，可能无法及时止盈 |
+
+---
+
+## 🔍 二、亏损原因分析
+
+### ✅ 1. **趋势判断滞后 + 止损设置不合理**
+
+- 策略使用了 EMA200 和价格背离作为主要信号
+- 在下行行情中：
+  - EMA200 反应慢，导致入场点位滞后
+  - 价格背离信号容易“假突破”，造成频繁误判
+- 当前 trailing stop 设置为 `0.237`，但在下跌行情中根本无法触发，反而被提前止损
+
+#### 📌 示例：
+
+```json
+"trailing_stop_positive": 0.237,
+"trailing_stop_positive_offset": 0.318,
+```
+
+这意味着只有当价格反弹到比最低价高出 23.7% 才会开始追踪，这在下跌趋势中几乎不可能实现。
+
+---
+
+### ✅ 2. **模型预测偏差 + 动态阈值失效**
+
+你的策略依赖 FreqAI 的 `&-price_value_divergence` 来判断买卖点，但在下行行情中：
+
+- 模型输出的 `labels_mean` 和 `labels_std` 偏移较大
+- 动态阈值计算逻辑过于简单，没有考虑市场状态变化
+- 导致买入信号过早触发，卖出信号延迟或缺失
+
+#### 📌 示例代码段：
+
+```python
+self.buy_threshold = labels_mean - k_buy * labels_std
+self.sell_threshold = labels_mean + k_sell * labels_std
+```
+
+在下行行情中，`labels_mean` 下降，但 `k_buy * labels_std` 仍保持不变，导致买点不准确。
+
+---
+
+### ✅ 3. **缺乏空头保护机制**
+
+你的策略是纯多头策略（`is_short: false`），但在下跌行情中：
+
+- 没有做空机制
+- 没有对冲手段
+- 没有根据大盘走势动态调整仓位或暂停交易
+
+#### 📌 回测日志显示：
+
+```log
+Market change: -6.96%
+```
+
+大盘整体下跌近 7%，而你的策略仍然坚持做多，导致系统性亏损。
+
+---
+
+### ✅ 4. **入场条件过于宽松**
+
+你的入场条件包括：
+
+```python
+cond1 = (dataframe["&-price_value_divergence"] < self.buy_threshold)
+cond2 = (dataframe["volume_z_score"] > 1.5)
+cond3 = (dataframe["rsi"] < 40)
+cond4 = (dataframe["close"] <= dataframe["bb_lowerband"])
+buy_condition = cond1 & cond2 & cond3 & cond4
+```
+
+这些条件在下跌行情中很容易满足，尤其是 RSI < 40 和 close 在布林下轨以下，容易形成“抄底陷阱”。
+
+---
+
+### ✅ 5. **出场逻辑单一 + ROI 不灵活**
+
+- ROI 设置为固定时间衰减曲线，但在下跌行情中需要更激进的止损
+- trailing stop 设置过高，无法有效锁定利润
+- sell_condition 中只有一个 RSI > 75 的条件，太弱
+
+---
+
+## 🚀 三、解决方案与优化建议
+
+### ✅ 1. **加入大盘趋势过滤器（Market Trend Filter）**
+
+你可以通过 BTC/USDT 或 ETH/USDT 的趋势来判断是否开启交易。
+
+#### 🔧 修改 `populate_indicators()`：
+
+```python
+def populate_indicators(self, dataframe: DataFrame, metadata: dict) -> DataFrame:
+    # 获取 BTC/USDT 的趋势
+    btc_df = self.dp.get_pair_dataframe("BTC/USDT", self.timeframe)
+    dataframe["btc_ema200"] = ta.EMA(btc_df, timeperiod=200)
+    dataframe["btc_trend"] = (btc_df["close"] > dataframe["btc_ema200"]).astype(int)
+
+    # 如果 BTC 处于下降趋势，则禁止交易
+    if dataframe["btc_trend"].iloc[-1] == 0:
+        logger.info(f"[{metadata['pair']}] BTC 处于下降趋势，暂停交易")
+        dataframe["enter_long"] = 0
+        return dataframe
+
+    # 否则继续执行原有逻辑...
+```
+
+---
+
+### ✅ 2. **动态调整 trailing stop 参数**
+
+在下跌行情中，适当降低 trailing stop 的距离，提高退出速度。
+
+#### 🔧 修改 `adjust_trade_position()`：
+
+```python
+def adjust_trade_position(self, trade: Trade, current_time: datetime,
+                        current_rate: float, current_profit: float,
+                        min_roi: Dict[float, float], max_profit: float):
+    hold_time = (current_time - trade.open_date_utc).total_seconds() / 60
+
+    if hold_time < 15:
+        return None
+
+    profit_ratio = (current_rate - trade.open_rate) / trade.open_rate
+
+    # 如果大盘处于下跌趋势，缩短 trailing stop 距离
+    market_trend = self.get_market_trend()
+    if market_trend == 'bear':
+        self.trailing_stop_distance = 0.005  # 更快退出
+        self.trailing_stop_start = 0.01
+    else:
+        self.trailing_stop_distance = 0.01
+        self.trailing_stop_start = 0.03
+
+    if profit_ratio >= self.trailing_stop_start and not self.trailing_stop_enabled:
+        self.trailing_stop_enabled = True
+        trade.adjust_max_rate(current_rate)
+
+    if self.trailing_stop_enabled:
+        max_rate = trade.max_rate
+        trailing_stop_price = max_rate * (1 - self.trailing_stop_distance)
+        if current_rate < trailing_stop_price:
+            return -1
+        trade.adjust_max_rate(current_rate)
+
+    if hold_time > 30:
+        return -1
+
+    return None
+```
+
+---
+
+### ✅ 3. **优化入场条件：增加趋势确认机制**
+
+避免在下跌趋势中盲目抄底。
+
+#### 🔧 修改 `populate_entry_trend()`：
+
+```python
+if "&-price_value_divergence" in dataframe.columns:
+    # 新增：只在价格高于短期均线时才允许入场
+    dataframe["ema20"] = ta.EMA(dataframe, timeperiod=20)
+    cond5 = (dataframe["close"] > dataframe["ema20"])
+
+    cond1 = (dataframe["&-price_value_divergence"] < self.buy_threshold)
+    cond2 = (dataframe["volume_z_score"] > 1.5)
+    cond3 = (dataframe["rsi"] < 40)
+    cond4 = (dataframe["close"] <= dataframe["bb_lowerband"])
+
+    buy_condition = cond1 & cond2 & cond3 & cond4 & cond5
+```
+
+---
+
+### ✅ 4. **动态调整 ROI 目标**
+
+在下跌行情中，ROI 应该更快衰减，甚至设为负值以强制止损。
+
+#### 🔧 修改 `generate_roi_table()`：
+
+```python
+def generate_roi_table(self, params: Dict) -> Dict[int, float]:
+    market_trend = self.get_market_trend()
+
+    if market_trend == 'bear':
+        roi_table = {
+            "0": -0.01,   # 强制止损
+            "5": 0.005,  # 极端情况下微利退出
+            "15": 0.01,
+            "30": 0.015
+        }
+    else:
+        roi_table = {
+            "0": params.get("roi_t0", 0.02),
+            "30": params.get("roi_t1", 0.01),
+            "60": params.get("roi_t2", 0.0),
+        }
+
+    return roi_table
+```
+
+---
+
+### ✅ 5. **加入大盘联动因子（Pair Correlation）**
+
+你可以利用 BTC/USDT 的波动率来动态调整其他币种的止损幅度。
+
+#### 🔧 示例逻辑：
+
+```python
+def custom_stoploss(self, pair: str, trade: Trade, current_time: datetime,
+                    current_rate: float, current_profit: float, **kwargs) -> float:
+
+    # 获取 BTC 波动率
+    btc_df = self.dp.get_pair_dataframe("BTC/USDT", self.timeframe)
+    volatility = (btc_df["close"].pct_change(24)).abs().mean()
+
+    base_stop = -0.05  # 基础止损
+    adjusted_stop = base_stop * (1 + volatility * 2)  # 波动越大，止损越紧
+
+    return max(adjusted_stop, -0.1)  # 最大不超过 -10%
+```
+
+---
+
+### ✅ 6. **加入“市场状态识别”模块**
+
+你可以使用 RSI + ADX + Volatility 综合判断市场状态。
+
+#### 🔧 示例逻辑：
+
+```python
+def get_market_trend(self, dataframe: DataFrame) -> str:
+    rsi = dataframe["rsi"].iloc[-1]
+    adx = dataframe["adx"].iloc[-1]
+    trend_strength = abs(dataframe["trend_strength"].iloc[-1])
+
+    if adx > 25 and trend_strength < -0.01:
+        return "bear"
+    elif adx > 25 and trend_strength > 0.01:
+        return "bull"
+    else:
+        return "sideways"
+```
+
+然后在 `populate_entry_trend()` 和 `custom_stoploss()` 中使用这个状态。
+
+---
+
+## ✅ 四、推荐配置修改（config.json）
+
+```json
+{
+  "stoploss": -0.1,
+  "minimal_roi": {
+    "0": -0.01,
+    "5": 0.005,
+    "15": 0.01,
+    "30": 0.015
+  },
+  "trailing_stop": true,
+  "trailing_stop_positive": 0.01,
+  "trailing_stop_positive_offset": 0.02,
+  "trailing_only_offset_is_reached": true
+}
+```
+
+---
+
+## ✅ 五、总结：亏损原因 & 对策一览表
+
+| 问题 | 原因 | 解决方案 |
+|------|------|----------|
+| 趋势判断滞后 | 使用 EMA200 + price divergence | 加入 BTC 趋势过滤器 |
+| 动态阈值失效 | 没有考虑市场状态 | 根据市场状态调整 buy/sell threshold |
+| 止损设置不合理 | trailing_stop_positive 过高 | 动态调整 trailing stop |
+| 入场条件宽松 | 容易抄底失败 | 加入 EMA20 过滤 |
+| 缺乏空头保护 | 没有做空机制 | 加入大盘趋势过滤器 |
+| ROI 设置不合理 | 固定衰减，未适应下跌行情 | 动态 ROI 表 |
+| 模型漂移 | labels_mean/std 偏移 | 加入 Z-score 标准化 |
+
+---
+
+🎯 **下一步建议：**
+
+你可以优先实现以下功能之一，我为你定制具体代码：
+
+| 功能 | 推荐指数 | 描述 |
+|------|----------|------|
+| BTC 趋势过滤器 | ⭐⭐⭐⭐⭐ | 防止在熊市中盲目做多 |
+| 动态 trailing stop | ⭐⭐⭐⭐ | 提高退出效率 |
+| 动态 ROI 表 | ⭐⭐⭐⭐ | 自适应不同市场环境 |
+| 市场状态识别 | ⭐⭐⭐ | 提升策略鲁棒性 |
+| 模型漂移检测 | ⭐⭐⭐ | 防止模型退化 |
+
+---
+
+如果你希望我帮你实现其中某一个功能，请告诉我你想优先实现哪个功能，我可以为你生成完整代码。
+
+------

chat/result5.md

@@ -0,0 +1,84 @@
+根据提供的日志、策略代码和配置文件，我将从以下几个方面进行排查分析：
+
+---
+
+## 🔍 一、FreqAI 是否正常训练与预测
+
+### ✅ 日志中的关键信息
+- **模型初始化成功**：
+  ```log
+  freqtrade.resolvers.strategy_resolver - INFO - Strategy using process_only_new_candles: True
+  ```
+- **特征工程完成**：
+  ```log
+  OKXRegressionStrategy - DEBUG - 目标初始化完成。DataFrame 形状：(43250, 475)
+  ```
+- **模型训练完成**：
+  ```log
+  BaseRegressionModel - INFO - -------------------- Done training OKB/USDT (219.77 secs) --------------------
+  ```
+
+### 📌 分析结论
+✅ FreqAI 模型确实完成了训练，并尝试进行了预测。
+
+---
+
+## ❗ 二、预测列未正确加载导致信号缺失
+
+### ⚠️ 日志中发现的问题
+```log
+OKXRegressionStrategy - WARNING - ⚠️ 缺少以下预测列：['pred_upper', 'pred_lower']
+```
+这说明在 `populate_indicators` 中尝试访问的 `pred_upper`, `pred_lower` 列不存在。
+
+### 📄 策略代码片段（问题所在）
+```python
+required_pred_cols = ['&-s_close', 'pred_upper', 'pred_lower']
+missing_pred_cols = [col for col in required_pred_cols if col not in dataframe.columns]
+
+if not missing_pred_cols:
+    logger.debug("✅ 预测列已就绪：")
+    logger.debug(dataframe[['&-s_close', 'pred_upper', 'pred_lower']].head().to_string())
+else:
+    logger.warning(f"⚠️ 缺少以下预测列：{missing_pred_cols}")
+```
+
+### 🧩 原因分析
+FreqAI 的预测结果默认列名可能不是 `pred_upper` 和 `pred_lower`，而是类似 `pred_upper_&-s_close` 这样的格式，尤其是当你使用了多个目标标签时。
+
+---
+
+## 🛠️ 解决方案：更新列名匹配逻辑
+
+### ✅ 修改建议
+将 `pred_upper` 和 `pred_lower` 替换为完整的列名模式，例如：
+```python
+'pred_upper_&-s_close'
+'pred_lower_&-s_close'
+```
+
+### 🔁 SEARCH/REPLACE 块如下：
+
+Filepath: freqtrade/templates/OKXRegressionStrategy.py
+Replace lines: 331-332
+```python
+required_pred_cols = ['&-s_close', 'pred_upper_&-s_close', 'pred_lower_&-s_close']
+missing_pred_cols = [col for col in required_pred_cols if col not in dataframe.columns]
+```
+
+Filepath: freqtrade/templates/OKXRegressionStrategy.py
+Replace lines: 338-338
+```python
+logger.warning(f"⚠️ 缺少以下预测列：{missing_pred_cols} | 当前可用列：{list(dataframe.columns)}")
+```
+
+---
+
+## 📈 三、交易信号生成是否依赖预测值？
+
+### 📉 日志显示没有生成有效信号
+```log
+OKXRegressionStrategy - DEBUG - 生成 0 个做多信号
+```
+
+### 🧾 策略代码片段（问题所在）

chat/result6.md

@@ -0,0 +1,278 @@
+
+
+- Datetime: 2025-05-06 10:24:16
+
+- Model: _qwen3/qwen-plus-latest
+
+- Selected files:
+  - neo-tree filesystem [1]
+  - result/backtest-result-2025-05-06_02-15-16.json
+  - result/backtest-result-2025-05-06_02-15-16.meta.json
+  - result/backtest-result-2025-05-06_02-15-16_OKXRegressionStrategy.py
+  - result/backtest-result-2025-05-06_02-15-16_config.json
+  - result/output_filted.log
+  - freqtrade/templates/OKXRegressionStrategy.py
+  - config_examples/config_freqai.okx.json
+
+> 当前策略OKXRegressionStrategy.py, 配置文件, 运行回测厚结果包含在log中, 根据其中的debug,info等信息, 结合下面列出的官方文档, 排查当前策略,配置代码是否工作正常符合预期, 如果不是,请给出分析的结论以及解决问题的 代码 块替换
+> 
+> 	自定义策略: https://www.freqtrade.io/en/stable/strategy-customization/
+> 	freqai 特征工程 : https://www.freqtrade.io/en/stable/freqai-feature-engineering/
+> 	freqai 参数表: https://www.freqtrade.io/en/stable/freqai-parameter-table/
+> 	freqai 配置: https://www.freqtrade.io/en/stable/freqai-configuration/ 
+> 	freaai 运行: https://www.freqtrade.io/en/stable/freqai-running/
+> 	freqai 开发手册: https://www.freqtrade.io/en/stable/freqai-developers/
+>   
+
+根据你提供的策略代码 (`OKXRegressionStrategy.py`)、配置文件 (`config_freqai.okx.json`) 和回测日志 (`output_filted.log`)，我进行了以下分析：
+
+---
+
+## 🔍 **问题诊断**
+
+### ✅ 1. 模型加载状态
+在 `OKXRegressionStrategy.py` 的 `__init__` 方法中，你添加了模型路径和存在性检查逻辑：
+```python
+model_file = f"{freqai_model_path}/{model_identifier}/cb_okb_*.pkl"
+if any(os.path.exists(f.replace('*', '1743465600')) for f in [model_file]):
+    logger.info("✅ 模型文件存在，准备加载")
+else:
+    logger.warning("⚠️ 模型文件未找到，请确认是否训练完成")
+```
+
+但在日志中出现如下警告：
+```
+2025-05-06 02:12:04,218 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Could not find model at /freqtrade/user_data/models/test175/sub-train-OKB_1743465600/cb_okb_1743465600
+```
+
+➡️ **结论：模型未成功加载**。  
+虽然策略尝试从指定路径加载模型，但实际路径 `/freqtrade/user_data/models/test175/sub-train-OKB_1743465600/cb_okb_1745971200.pkl` 并不存在或未被正确生成。
+
+---
+
+### ❌ 2. 预测列缺失
+
+在 `populate_indicators()` 中有如下逻辑：
+```python
+required_pred_cols = ['pred_upper', 'pred_lower', '&-s_close', '&-roi_0_pred']
+missing_pred_cols = [col for col in required_pred_cols if col not in dataframe.columns]
+if missing_pred_cols:
+    logger.warning(f"⚠️ 缺失预测列：{missing_pred_cols}，尝试回退默认值")
+    for col in missing_pred_cols:
+        dataframe[col] = np.nan
+```
+
+日志中也出现了该警告：
+```
+2025-05-06 02:15:15,695 - OKXRegressionStrategy - WARNING - ⚠️ 缺失预测列：['pred_upper', 'pred_lower']，尝试回退默认值
+```
+
+➡️ **结论：模型未输出关键预测列（如 pred_upper, pred_lower）**。  
+这导致交易信号无法正常触发，因为 `enter_long` 是基于 `&-s_close` 和 `pred_upper` 的组合判断的。
+
+---
+
+### 📉 3. 回测结果表现异常
+
+回测结果显示：
+```
+Avg Profit %: -0.26%
+Tot Profit USDT: -3.811
+Winrate: 47.3%
+Profit Factor: 0.48
+```
+
+➡️ **结论：策略整体表现不佳，胜率低于 50%，且盈利因子小于 1，说明策略不具备正期望收益。**
+
+---
+
+## 🧰 **根本原因分析**
+
+| 原因 | 描述 |
+|------|------|
+| 模型未保存/加载 | 日志显示模型路径下没有 `.pkl` 文件，说明 FreqAI 可能未成功保存模型 |
+| 特征工程与目标列不匹配 | 策略定义了多个特征列，但模型可能未输出对应预测列（如 `pred_upper`, `pred_lower`） |
+| 预测列缺失导致信号失效 | `enter_long` 依赖于 `&-s_close > 0.005` 或 `rsi < 30`，但由于预测列缺失，信号未能有效生成 |
+
+---
+
+## 🛠️ **解决方案**
+
+### ✅ 1. 确保模型训练并保存
+
+在 `config_freqai.okx.json` 中确保：
+```json
+"save_backtesting_prediction": true,
+"purge_old_models": false,
+"load_trained_model": true
+```
+
+同时，在命令行运行时加上 `--freqaimodel-path` 参数以指定模型路径：
+```bash
+freqtrade backtesting --strategy OKXRegressionStrategy --config config_freqai.okx.json --freqaimodel-path user_data/models/
+```
+
+---
+
+### ✅ 2. 在策略中强制等待模型加载
+
+修改 `set_freqai_targets()` 方法，确保模型已加载后再继续执行：
+```python
+def set_freqai_targets(self, dataframe: DataFrame, metadata: Dict, **kwargs) -> DataFrame:
+    # 等待模型加载
+    while not hasattr(self, "freqai") or self.freqai is None:
+        time.sleep(1)
+
+    # 正常设置目标列
+    dataframe["&-s_close"] = (dataframe["close"].shift(-self.label_period_candles) - dataframe["close"]) / dataframe["close"]
+    ...
+```
+
+---
+
+### ✅ 3. 添加模型训练失败处理逻辑
+
+在 `fit()` 方法中增加重试机制和错误提示：
+```python
+def fit(self, data_dictionary: Dict, metadata: Dict, **kwargs) -> None:
+    try:
+        if not hasattr(self, 'model') or self.model is None:
+            self.model = XGBRegressor(**self.freqai_config["model_training_parameters"])
+        self.freqai.fit(data_dictionary, metadata, **kwargs)
+    except Exception as e:
+        logger.error(f"FreqAI 模型训练失败：{str(e)}")
+        raise RuntimeError("模型训练失败，请检查数据完整性或重新训练模型")
+```
+
+---
+
+### ✅ 4. 强制填充预测列（用于调试）
+
+在 `populate_indicators()` 中加入模拟预测列逻辑，防止空列影响信号生成：
+```python
+# 如果预测列缺失，用随机数模拟
+if 'pred_upper' not in dataframe.columns:
+    dataframe['pred_upper'] = dataframe['close'].rolling(window=10).mean() * 1.01
+if 'pred_lower' not in dataframe.columns:
+    dataframe['pred_lower'] = dataframe['close'].rolling(window=10).mean() * 0.99
+```
+
+---
+
+## ✅ 最终建议
+
+| 动作 | 建议 |
+|------|------|
+| ✅ 模型训练 | 使用 `freqtrade train` 显式训练模型，并确认模型保存到 `user_data/models/` |
+| ✅ 模型加载 | 修改策略逻辑，确保模型加载完成再进行指标计算 |
+| ✅ 数据完整性 | 在 `set_freqai_targets()` 中验证所有目标列是否存在 |
+| ✅ 模拟预测列 | 在 `populate_indicators()` 中加入模拟预测列，防止空列导致信号丢失 |
+| ✅ 调整参数 | 尝试不同 `n_estimators`, `learning_rate`, `label_period_candles` 提高模型表现 |
+
+---
+
+## 📦 SEARCH/REPLACE 块
+
+Filepath: freqtrade/templates/OKXRegressionStrategy.py
+Replace lines: 0-0
+```python
+    def set_freqai_targets(self, dataframe: DataFrame, metadata: Dict, **kwargs) -> DataFrame:
+        """
+        设置回归模型的目标变量，为不同币对设置动态止损和ROI阈值。
+        输入：dataframe（K线数据：close, high, low），metadata（交易对信息），config（FreqAI配置）
+        输出：更新后的dataframe，包含目标标签
+        """
+        # 等待模型加载完成
+        import time
+        while not hasattr(self, "freqai") or self.freqai is None:
+            time.sleep(1)
+
+        # 获取配置参数
+        label_period = self.freqai_config["feature_parameters"]["label_period_candles"]  # 标签预测周期（如5分钟K线的N根）
+        pair = metadata["pair"]  # 当前交易对（如DOGE/USDT）
+
+        # 计算未来价格变化率（现有逻辑）
+        dataframe["&-s_close"] = (dataframe["close"].shift(-label_period) - dataframe["close"]) / dataframe["close"]
+
+        # 计算不同时间窗口的ROI（现有逻辑）
+        for minutes in [0, 15, 30]:
+            candles = int(minutes / 5)  # 假设5分钟K线
+            if candles > 0:
+                dataframe[f"&-roi_{minutes}"] = (dataframe["close"].shift(-candles) - dataframe["close"]) / dataframe["close"]
+            else:
+                dataframe[f"&-roi_{minutes}"] = 0.0
+
+        # 计算市场状态指标：ADX（14周期，与label_period_candles对齐）
+        dataframe["adx"] = ta.ADX(dataframe["high"], dataframe["low"], dataframe["close"], timeperiod=14)
+
+        # 定义币对特定的ADX阈值和止损/ROI范围
+        pair_thresholds = {
+            "DOGE/USDT": {
+                "adx_trend": 20,
+                "adx_oscillation": 15,
+                "stoploss_trend": -0.08,
+                "stoploss_oscillation": -0.04,
+                "stoploss_mid": -0.06,
+                "roi_trend": 0.06,
+                "roi_oscillation": 0.025,
+                "roi_mid": 0.04
+            },
+            "BTC/USDT": {
+                "adx_trend": 25,
+                "adx_oscillation": 20,
+                "stoploss_trend": -0.03,
+                "stoploss_oscillation": -0.015,
+                "stoploss_mid": -0.02,
+                "roi_trend": 0.03,
+                "roi_oscillation": 0.015,
+                "roi_mid": 0.02
+            },
+            "SOL/USDT": {
+                "adx_trend": 22,
+                "adx_oscillation": 18,
+                "stoploss_trend": -0.06,
+                "stoploss_oscillation": -0.03,
+                "stoploss_mid": -0.045,
+                "roi_trend": 0.045,
+                "roi_oscillation": 0.02,
+                "roi_mid": 0.03
+            },
+            "XRP/USDT": {
+                "adx_trend": 22,
+                "adx_oscillation": 18,
+                "stoploss_trend": -0.06,
+                "stoploss_oscillation": -0.03,
+                "stoploss_mid": -0.045,
+                "roi_trend": 0.045,
+                "roi_oscillation": 0.02,
+                "roi_mid": 0.03
+            },
+            "OKB/USDT": {
+                "adx_trend": 18,
+                "adx_oscillation": 12,
+                "stoploss_trend": -0.10,
+                "stoploss_oscillation": -0.06,
+                "stoploss_mid": -0.08,
+                "roi_trend": 0.07,
+                "roi_oscillation": 0.04,
+                "roi_mid": 0.05
+            }
+        }
+
+
+        # 动态化 &-roi_0_pred（基于市场状态和币对）
+        dataframe["&-roi_0_pred"] = 0.0
+        for index, row in dataframe.iterrows():
+            thresholds = pair_thresholds.get(pair, {})
+            if not thresholds:
+                continue
+            adx_value = row["adx"]
+            if adx_value > thresholds["adx_trend"]:  # 趋势市场
+                dataframe.at[index, "&-roi_0_pred"] = thresholds["roi_trend"]  # 较高ROI目标
+            elif adx_value < thresholds["adx_oscillation"]:  # 震荡市场
+                dataframe.at[index, "&-roi_0_pred"] = thresholds["roi_oscillation"]  # 较低ROI目标
+            else:
+                # 中性市场：使用中间值
+                dataframe.at[index, "&-roi_0_pred"] = thresholds["roi_mid"]
+
+        return dataframe

chat/result7.md

@@ -0,0 +1,244 @@
+根据当前提供的策略文件 `OKXRegressionStrategy.py`、配置文件 `config_freqai.okx.json` 和回测日志 `output_filted.log`，结合 Freqtrade 官方文档和实际运行逻辑，我进行了以下分析：
+
+---
+
+## ✅ **1. 策略是否正常工作？**
+
+### ❌ **结论：策略未按预期工作。**
+
+尽管代码结构基本符合 FreqAI 的要求，但存在几个关键问题导致模型预测未被正确使用，最终交易信号仅依赖 RSI。
+
+---
+
+## 🧨 **2. 主要问题分析**
+
+### 🔴 2.1 模型未加载成功
+
+**日志证据：**
+```
+2025-05-06 03:03:03,093 - OKXRegressionStrategy - WARNING - ⚠️ 模型文件未找到，请确认是否训练完成
+2025-05-06 03:04:29,843 - freqtrade.plot.plotting - INFO - Stored plot as /freqtrade/user_data/models/test175/sub-train-OKB_1743465600/cb_okb_1743465600--s_close.html
+2025-05-06 03:05:53,910 - OKXRegressionStrategy - WARNING - ⚠️ pred_upper 列全为 NaN，可能模型未加载成功
+```
+
+**原因：**
+- 日志显示模型路径 `/freqtrade/user_data/models/test175/` 下没有 `.pkl` 文件；
+- 虽然回测过程中重新训练了模型（生成了 HTML 图表），但这些模型未被保存或加载到策略中用于预测；
+- 回测结束后，`pred_upper`, `pred_lower` 列仍为 `NaN`，说明模型输出未生效。
+
+**官方文档参考：**
+- [FreqAI Running](https://www.freqtrade.io/en/stable/freqai-running/)：需确保模型在训练后保存，并在回测时通过 `load_trained_model: true` 加载；
+- [FreqAI Configuration](https://www.freqtrade.io/en/stable/freqai-configuration/)：检查 `model_filename` 是否指向正确的模型路径。
+
+---
+
+### 🔴 2.2 预测列缺失或无效
+
+**日志证据：**
+```
+2025-05-06 03:05:53,910 - OKXRegressionStrategy - WARNING - ⚠️ pred_upper 列全为 NaN，可能模型未加载成功
+2025-05-06 03:05:53,917 - OKXRegressionStrategy - DEBUG -    &-s_close  pred_upper  pred_lower
+0        0.0         NaN         NaN
+1        0.0         NaN         NaN
+2        0.0         NaN         NaN
+3        0.0         NaN         NaN
+4        0.0         NaN         NaN
+```
+
+**后果：**
+- `populate_entry_trend()` 中的 `valid_prediction = (~dataframe['&-s_close'].isna()) & (dataframe['&-s_close'] > 0.005)` 实际上只依赖常量 `0.0`；
+- 所有做多信号仅由 RSI < 30 触发，与模型无关。
+
+**官方文档参考：**
+- [FreqAI Feature Engineering](https://www.freqtrade.io/en/stable/freqai-feature-engineering/)：必须确保目标列如 `&-s_close` 被正确识别并用于训练；
+- [Custom Strategy](https://www.freqtrade.io/en/stable/strategy-customization/)：交易信号应基于有效预测值。
+
+---
+
+### 🔴 2.3 模型训练未持久化
+
+**日志证据：**
+```
+2025-05-06 03:04:43,510 - freqtrade.freqai.freqai_interface - INFO - Could not find model at /freqtrade/user_data/models/test175/sub-train-OKB_1744329600/cb_okb_1744329600
+```
+
+**后果：**
+- 每次回测都重新训练模型，且未保存模型供后续使用；
+- 导致无法复用已有模型进行稳定回测。
+
+**建议：**
+- 使用 `freqtrade train` 显式训练模型并保存；
+- 在回测前确保模型已存在；
+- 设置 `"purge_old_models": false` 保留历史模型。
+
+---
+
+## 🛠️ **3. 解决方案与代码修复**
+
+### ✅ **3.1 修改策略类名以匹配文件名**
+
+**问题：**
+- 当前策略类名为 `OKXRegressionStrategy`，但文件名也为 `OKXRegressionStrategy.py`；
+- 若从模板导入，需确保类名一致。
+
+**修改建议：**
+
+Filepath: freqtrade/templates/OKXRegressionStrategy.py
+Replace lines: 18-18
+```python
+class OKXRegressionStrategy(IStrategy):
+```
+
+✅ 无需更改，当前命名已正确。
+
+---
+
+### ✅ **3.2 强制启用模型保存与加载**
+
+**问题：**
+- 回测时未加载已有模型；
+- 每次都重新训练，影响可重复性。
+
+**修改建议：**
+
+Filepath: config_examples/config_freqai.okx.json
+Replace lines: 62-68
+```json
+"freqai": {
+    "enabled": true,
+    "identifier": "test175",
+    "freqaimodel": "XGBoostRegressor",
+    "save_backtesting_prediction": true,
+    "purge_old_models": false,
+    "load_trained_model": true,
+    "model_filename": "/freqtrade/user_data/models/test175/cb_okb_*.pkl"
+```
+
+📌 **说明：**
+- 增加 `model_filename` 字段，明确指定模型路径；
+- 确保模型能被正确加载用于预测。
+
+---
+
+### ✅ **3.3 补充模型训练命令以生成 `.pkl` 文件**
+
+**问题：**
+- 当前直接运行回测会触发训练，但不会持久化模型；
+- 应先显式训练模型再进行回测。
+
+**解决方案：**
+
+```bash
+# 先训练模型
+docker-compose run --rm freqtrade train --config config_examples/config_freqai.okx.json --freqaimodel XGBoostRegressor --strategy OKXRegressionStrategy
+
+# 再运行回测
+docker-compose run --rm freqtrade backtesting --config config_examples/config_freqai.okx.json --freqaimodel XGBoostRegressor --strategy OKXRegressionStrategy
+```
+
+📌 **说明：**
+- 训练完成后会在 `/freqtrade/user_data/models/test175/` 下生成 `.pkl` 模型文件；
+- 回测时将加载该模型进行预测。
+
+---
+
+### ✅ **3.4 确保 `set_freqai_targets()` 正确设置目标列**
+
+**问题：**
+- 目标列 `&-s_close`, `&-roi_0_pred` 已定义；
+- 但在某些情况下未被 FreqAI 正确识别。
+
+**修改建议：**
+
+Filepath: freqtrade/templates/OKXRegressionStrategy.py
+Replace lines: 0-0
+```python
+def set_freqai_targets(self, dataframe: DataFrame, metadata: Dict) -> DataFrame:
+    # 设置回归目标
+    label_period = self.freqai_config["feature_parameters"]["label_period_candles"]
+    dataframe["&-s_close"] = (dataframe["close"].shift(-label_period) - dataframe["close"]) / dataframe["close"]
+
+    # 设置动态 ROI 和止损阈值
+    dataframe["&-roi_0_pred"] = 0.03  # 默认值，防止空值
+    dataframe["&-stoploss_pred"] = -0.05
+
+    return dataframe
+```
+
+📌 **说明：**
+- 提供默认值避免空列；
+- 确保列名与 `feature_engineering_expand_all()` 中的特征一致；
+- 可进一步扩展为基于 ADX 的动态设定。
+
+---
+
+### ✅ **3.5 确保 `populate_entry_trend()` 使用模型预测值**
+
+**问题：**
+- 当前 `valid_prediction` 条件过于宽松，仅判断 `&-s_close > 0.005`；
+- 若模型未加载，此条件始终成立，导致信号仅由 RSI 控制。
+
+**修改建议：**
+
+Filepath: freqtrade/templates/OKXRegressionStrategy.py
+Replace lines: 365-365
+```python
+valid_prediction = (~dataframe['&-s_close'].isna()) & (dataframe['&-s_close'] > 0.005) & (dataframe['&-s_close'].abs() > 0.001)
+```
+
+📌 **说明：**
+- 添加 `abs() > 0.001` 过滤掉恒定为零的预测值；
+- 确保只有当模型真正输出有意义数值时才生成信号。
+
+---
+
+## 📊 **4. 验证建议**
+
+请执行以下操作验证模型是否加载成功并参与预测：
+
+1. **运行训练命令生成模型：**
+   ```bash
+   docker-compose run --rm freqtrade train --config config_examples/config_freqai.okx.json --freqaimodel XGBoostRegressor --strategy OKXRegressionStrategy
+   ```
+
+2. **查看模型是否生成：**
+   ```bash
+   ls -la /home/ubuntu/freqtrade/user_data/models/test175/
+   ```
+   应看到类似 `cb_okb_*.pkl` 的模型文件。
+
+3. **运行回测并检查预测列：**
+   ```bash
+   docker-compose run --rm freqtrade backtesting --config config_examples/config_freqai.okx.json --freqaimodel XGBoostRegressor --strategy OKXRegressionStrategy
+   ```
+
+4. **打开 `result/backtest-result-*.json` 查看预测列：**
+   - 检查 `&-s_close`, `pred_upper`, `pred_lower` 是否随时间变化；
+   - 若仍为恒定值或全为 NaN，则说明模型仍未加载成功。
+
+---
+
+## ✅ **5. 总结**
+
+| 问题 | 是否解决 | 说明 |
+|------|----------|------|
+| 模型未加载 | ❌ | 缺少 `.pkl` 模型文件 |
+| 预测列为 NaN | ❌ | 模型未加载导致预测失败 |
+| 交易信号仅由 RSI 触发 | ❌ | `&-s_close` 恒定为 0 |
+| 模型未持久化 | ❌ | 每次回测都重新训练 |
+| 策略类名不一致 | ✅ | 类名与文件名一致 |
+
+---
+
+## 📌 **下一步建议**
+
+请优先执行以下任意一项操作以继续诊断：
+
+1. **提供 `result/backtest-result-*.json` 中部分预测数据片段；**
+2. **提供 `output_filted.log` 中关于模型加载和预测值的关键日志；**
+3. **告知是否可以运行新的训练 + 回测流程并启用 `save_backtesting_prediction: true`。**
+
+这将帮助我们进一步确认模型是否加载成功以及预测值是否合理。
+
+**Generation complete!** Please review the code suggestions above.
+

chat/result8.md

@@ -0,0 +1,6 @@
+### 🛠️ 可能的改进方向
+
+如果你希望进一步优化这个替代逻辑，可以考虑：
+1. **引入更多技术指标**（如 MACD、布林带等）来增强模拟预测的智能性。
+2. **使用简单机器学习模型**（如线性回归）基于历史 RSI 和价格走势训练一个轻量级替代模型。
+3. **动态调整窗口大小**（rolling window），根据市场波动率自动调节平滑程度。

chat/roadmap.md

@@ -0,0 +1,68 @@
+> 根据当前问题, 制定一个优化路线图, 每次只完成一个功能点改动, 路线图只包含要优化的问题,和解决思路,不包含代码,请输出此路线图
+> 
+
+### ✅ **OKXRegressionStrategy 优化路线图**
+
+| 步骤 | 待优化问题 | 解决思路 |
+|------|------------|----------|
+| 1 | **胜率低（33%–50%）** | 引入模型预测置信度过滤，仅在 `pred_upper - pred_lower > 0.02` 时允许交易，减少低质量信号。 |
+| 2 | **退出逻辑单一（仅依赖 exit_signal）** | 增加回调止损（Chandelier Exit）和动态 ROI 结合使用，提升止盈/止损效率。 |
+| 3 | **固定 ATR multiplier 不适应波动变化** | 使用历史 ATR 百分位动态调整 multiplier：高波动 → 更紧止损；低波动 → 更松止损。 |
+| 4 | **缺乏市场状态识别（趋势 vs 震荡）** | 利用 ADX 判断市场状态，并据此调整 exit_signal 条件阈值，提高趋势跟踪能力。 |
+| 5 | **特征工程维度不足** | 添加 BB Width、Volume Change、Trend Diff 等高阶特征，增强模型对价格行为的理解。 |
+| 6 | **模型训练未启用交叉验证（CV）** | 启用 TimeSeriesSplit CV，提升模型稳定性与泛化能力。 |
+| 7 | **回撤过大（最大达 7.341 USDT）** | 引入回调止损逻辑（callback stop loss），防止趋势中途反转造成大幅亏损。 |
+| 8 | **无动态仓位管理** | 根据账户风险暴露和波动率动态调整 stake_amount，降低极端行情下的单笔损失。 |
+
+---
+
+📌 **说明**：
+- 每次只完成一个功能点改动，确保每次 backtest 变量唯一。
+- 每个改动后运行 backtest 并记录关键指标（胜率、平均盈利、最大回撤等）。
+- 若某项改动效果显著，可继续深入调参；若无效，则跳过或尝试替代方案。
+
+是否需要我开始第一步，即“引入模型预测置信度过滤”？
+
+-------
+
+- Datetime: 2025-05-05 17:37:56
+
+- Model: _qwen3/qwen-plus-latest
+
+- Selected files:
+  - chat/report.md
+  - chat/report2.md
+  - freqtrade/templates/OKXRegressionStrategy.py
+
+> 根据路线图的内容描述,当前代码中的部分功能点是否要被删掉,如果需要,请在路线图中明确说明,否则新版代码将新旧混杂,难以维护
+
+你提出了一个非常关键的问题：**当前代码中确实存在一些“新旧混杂”的功能点，如果不清理，将影响策略的可维护性和逻辑一致性。**
+
+---
+
+## ✅ **确认需要删除的功能点（已在路线图中覆盖）**
+
+| 功能点 | 是否保留 | 说明 |
+|--------|----------|------|
+| `exit_signal` 基于固定 RSI 和预测值下降 | ❌ 删除 | 已被市场状态 + Chandelier Exit 替代 |
+| `_dynamic_stop_loss` 方法 | ❌ 删除 | 已被 `custom_stoploss` + ATR 百分位替代 |
+| `populate_exit_trend` 中硬编码 exit_signal | ❌ 删除 | 已由回调止损和动态 ROI 替代 |
+| `check_data_freshness()` 简化版 | ❌ 删除 | 实际运行中依赖 Freqtrade 数据源质量，无需额外检查 |
+| `confirm_trade_entry()` 中 OKX 数据验证 | ❌ 删除 | 同上，且 OKX 不是唯一数据来源 |
+| `fetch_okx_order_book()` 注释掉的订单簿逻辑 | ✅ 保留（注释） | 可作为未来扩展参考 |
+| `feature_engineering_expand_all` 中冗余缓存逻辑 | ✅ 保留 | 缓存机制仍有效，提升性能 |
+| `fit()` 方法中的手动模型初始化 | ✅ 保留 | 与 FreqAI 兼容性良好 |
+
+
+### ✅ **OKXRegressionStrategy 优化路线图**
+
+| 步骤 | 待优化问题 | 解决思路 | 是否需要删除旧代码 |
+|------|------------|----------|------------------|
+| 1 | **胜率低（33%–50%）** | 引入模型预测置信度过滤，仅在 `pred_upper - pred_lower > 0.02` 时允许交易，减少低质量信号。 | 否 |
+| 2 | **退出逻辑单一（仅依赖 exit_signal）** | 增加回调止损（Chandelier Exit）和动态 ROI 结合使用，提升止盈/止损效率。 | 是（`exit_signal` 已被替代） |
+| 3 | **固定 ATR multiplier 不适应波动变化** | 使用历史 ATR 百分位动态调整 multiplier：高波动 → 更紧止损；低波动 → 更松止损。 | 是（`_dynamic_stop_loss` 方法已废弃） |
+| 4 | **缺乏市场状态识别（趋势 vs 震荡）** | 利用 ADX 判断市场状态，并据此调整 exit_signal 条件阈值，提高趋势跟踪能力。 | 否 |
+| 5 | **特征工程维度不足** | 添加 BB Width、Volume Change、Trend Diff 等高阶特征，增强模型对价格行为的理解。 | 否 |
+| 6 | **模型训练未启用交叉验证（CV）** | 启用 TimeSeriesSplit CV，提升模型稳定性与泛化能力。 | 否 |
+| 7 | **回撤过大（最大达 7.341 USDT）** | 引入回调止损逻辑（callback stop loss），防止趋势中途反转造成大幅亏损。 | 是（原 `populate_exit_trend` 中硬编码 exit_signal 被替代） |
+| 8 | **无动态仓位管理** | 根据账户风险暴露和波动率动态调整 stake_amount，降低极端行情下的单笔损失。 | 否 |

code.txt

@@ -0,0 +1,101 @@
+ FUNCTION set_freqai_targets(dataframe, metadata, config):
+     """
+     设置回归模型的目标变量，为不同币对设置动态止损和ROI阈值。
+     输入：dataframe（K线数据：close, high, low），metadata（交易对信息），config（FreqAI配置）
+     输出：更新后的dataframe，包含目标标签
+     """
+     # 获取配置参数
+     label_period = `config["feature_parameters"]`["label_period_candles"]  # 标签预测周期（如5分钟K线的N根）
+     pair = `metadata["pair"]`  # 当前交易对（如DOGE/USDT）
+ 
+     # 计算未来价格变化率（现有逻辑）
+     SET `dataframe["&-s_close"]` = (`dataframe["close"]`.shift(-label_period) - `dataframe["close"]`) / `dataframe["close"]`
+ 
+     # 计算不同时间窗口的ROI（现有逻辑）
+     FOR minutes IN [0, 15, 30]:
+         candles = INTEGER(minutes / 5)  # 假设5分钟K线
+         IF candles > 0:
+             SET `dataframe["&-roi_{minutes}"]` = (`dataframe["close"]`.shift(-candles) - `dataframe["close"]`) / `dataframe["close"]`
+         ELSE:
+             SET `dataframe["&-roi_{minutes}"]` = 0.0
+ 
+     # 计算市场状态指标：ADX（14周期，与label_period_candles对齐）
+     SET `dataframe["adx"]` = CALCULATE_ADX(`dataframe["high"]`, `dataframe["low"]`, `dataframe["close"]`, period=14)
+ 
+     # 定义币对特定的ADX阈值和止损/ROI范围
+     INITIALIZE pair_thresholds = {
+         "DOGE/USDT": {
+             "adx_trend": 20,          # 趋势市场ADX阈值
+             "adx_oscillation": 15,    # 震荡市场ADX阈值
+             "stoploss_trend": -0.08,  # 趋势市场止损：-8%
+             "stoploss_oscillation": -0.04,  # 震荡市场止损：-4%
+             "stoploss_mid": -0.06,    # 中间状态止损：-6%
+             "roi_trend": 0.06,        # 趋势市场ROI：6%
+             "roi_oscillation": 0.025, # 震荡市场ROI：2.5%
+             "roi_mid": 0.04           # 中间状态ROI：4%
+         },
+         "BTC/USDT": {
+             "adx_trend": 25,
+             "adx_oscillation": 20,
+             "stoploss_trend": -0.03,
+             "stoploss_oscillation": -0.015,
+             "stoploss_mid": -0.02,
+             "roi_trend": 0.03,
+             "roi_oscillation": 0.015,
+             "roi_mid": 0.02
+         },
+         "SOL/USDT": {
+             "adx_trend": 22,
+             "adx_oscillation": 18,
+             "stoploss_trend": -0.06,
+             "stoploss_oscillation": -0.03,
+             "stoploss_mid": -0.045,
+             "roi_trend": 0.045,
+             "roi_oscillation": 0.02,
+             "roi_mid": 0.03
+         },
+         "XRP/USDT": {
+             "adx_trend": 22,
+             "adx_oscillation": 18,
+             "stoploss_trend": -0.06,
+             "stoploss_oscillation": -0.03,
+             "stoploss_mid": -0.045,
+             "roi_trend": 0.045,
+             "roi_oscillation": 0.02,
+             "roi_mid": 0.03
+         }
+     }
+ 
+     # 动态化 &-stoploss_pred（基于市场状态和币对）
+     INITIALIZE `dataframe["&-stoploss_pred"]` = 0.0
+     FOR EACH row IN dataframe:
+         thresholds = `pair_thresholds[pair]`  # 获取当前币对的阈值
+         IF `row["adx"]` > `thresholds["adx_trend"]`:  # 趋势市场
+             SET `row["&-stoploss_pred"]` = `thresholds["stoploss_trend"]`  # 宽松止损
+         ELSE IF `row["adx"]` < `thresholds["adx_oscillation"]`:  # 震荡市场
+             SET `row["&-stoploss_pred"]` = `thresholds["stoploss_oscillation"]`  # 严格止损
+         ELSE:  # 中间状态
+             SET `row["&-stoploss_pred"]` = `thresholds["stoploss_mid"]`  # 中等止损
+         # 风险控制：设置止损下限
+         IF `row["&-stoploss_pred"]` < -0.10:  # 防止过大止损
+             SET `row["&-stoploss_pred"]` = -0.10
+ 
+     # 动态化 &-roi_0_pred（基于市场趋势和币对）
+     INITIALIZE `dataframe["&-roi_0_pred"]` = 0.0
+     FOR EACH row IN dataframe:
+         thresholds = `pair_thresholds[pair]`  # 获取当前币对的阈值
+         IF `row["adx"]` > `thresholds["adx_trend"]`:  # 强趋势市场
+             SET `row["&-roi_0_pred"]` = `thresholds["roi_trend"]`  # 高ROI
+         ELSE IF `row["adx"]` < `thresholds["adx_oscillation"]`:  # 震荡市场
+             SET `row["&-roi_0_pred"]` = `thresholds["roi_oscillation"]`  # 低ROI
+         ELSE:  # 中间状态
+             SET `row["&-roi_0_pred"]` = `thresholds["roi_mid"]`  # 中等ROI
+         # 风险控制：设置ROI上限
+         IF `row["&-roi_0_pred"]` > 0.10:  # 防止过高目标
+             SET `row["&-roi_0_pred"]` = 0.10
+ 
+     # 计算RSI预测（现有逻辑）
+     SET `dataframe["&-buy_rsi_pred"]` = CALCULATE_RSI(`dataframe["close"]`, period=14).ROLLING_MEAN(window=20)
+ 
+     RETURN dataframe
+

config_examples/basic.json

@@ -5,6 +5,7 @@
     "max_open_trades": 4,
     "stake_currency": "USDT",
     "stake_amount": 150,
+    "startup_candle_count": 30,
     "tradable_balance_ratio": 1,
     "fiat_display_currency": "USD",
     "dry_run": true,
@@ -62,49 +63,44 @@
     "freqai": {
         "enabled": true,
         "data_kitchen": {
-            "fillna": "ffill"
+            "fillna": "ffill",
+            "feature_parameters": {
+                "DI_threshold": 0.9,
+                "weight_factor": 0.9
+            }
         },
-        "freqaimodel": "CatboostClassifier",
+        "freqaimodel": "XGBoostRegressor",
         "purge_old_models": 2,
-        "train_period_days": 15,
-        "identifier": "test62",
+        "identifier": "test175",
         "train_period_days": 30,
         "backtest_period_days": 10,
         "live_retrain_hours": 0,
         "feature_selection": {
-          "method": "recursive_elimination"
+            "method": "recursive_elimination",
+            "threshold": 0.01
         },
         "feature_parameters": {
-            "include_timeframes": [
-                "3m",
-                "15m",
-                "1h"
-            ],
-            "include_corr_pairlist": [
-                "BTC/USDT",
-                "SOL/USDT"
-            ],
-            "label_period_candles": 20,
-            "include_shifted_candles": 2,
-            "DI_threshold": 0.9,
-            "weight_factor": 0.9,
-            "principal_component_analysis": false,
-            "use_SVM_to_remove_outliers": false,
-            "indicator_periods_candles": [
-                10,
-                20,
-                50
-            ],
-            "plot_feature_importances": 0
+            "include_timeframes": ["5m", "1h"],
+            "include_corr_pairlist": ["BTC/USDT"],
+            "label_period_candles": 10,
+            "use_SVM_to_remove_outliers": true,
+            "principal_component_analysis": true
         },
+        "model": "XGBoostRegressor",
         "data_split_parameters": {
-            "test_size": 0.2
+            "test_size": 0.2,
+            "shuffle": true,
+            "random_state": 42
         },
         "model_training_parameters": {
-            "n_estimators": 100,
+            "n_estimators": 200,
             "learning_rate": 0.05,
             "max_depth": 5,
-            "num_leaves": 31
+            "subsample": 0.8,
+            "colsample_bytree": 0.8,
+            "objective": "reg:squarederror",
+            "eval_metric": "rmse",
+            "early_stopping_rounds": 50
         }
     },
     "api_server": {

config_examples/config.json

@@ -0,0 +1,77 @@
+{
+    "exchange": {
+        "name": "okx",
+        "key": "your_api_key",
+        "secret": "your_api_secret",
+        "ccxt_config": {
+            "enableRateLimit": true,
+            "defaultType": "spot",
+            "rateLimit": 2000,
+            "timeout": 30000
+        },
+        "pair_whitelist": ["BTC/USDT",  "TON/USDT", "DOT/USDT", "XRP/USDT",  "OKB/USDT", "SOL/USDT", "DOGE/USDT", "WCT/USDT", "TRUMP/USDT", "SUI/USDT", "PEPE/USDT", "TRB/USDT", "MASK/USDT", "UNI/USDT", "KAITO/USDT"]
+    },
+    "pairlists": [
+        {
+            "method": "StaticPairList"
+        }
+    ],
+    "timeframe": "15m",
+    "strategy": "NostalgiaForInfinityV8",
+    "freqai": {
+        "enabled": true,
+        "freqaimodel": "LightGBMRegressor",
+        "train_period_days": 7,
+        "backtest_period_days": 0.2,
+        "feature_parameters": {
+            "include_timeframes": ["15m"],
+            "include_corr_pairlist": ["OKB/USDT", "SOL/USDT"],
+            "label_period_candles": 4,
+            "include_shifted_candles": 0,
+            "indicator_periods_candles": [5, 10, 20],
+            "include_default_features": ["close", "open", "high", "low", "volume", "rsi", "ema", "sma"],
+            "DI_threshold": 0.3
+        },
+        "data_split_parameters": {
+            "test_size": 0.2,
+            "shuffle": false
+        },
+        "model_training_parameters": {
+            "n_estimators": 100,
+            "learning_rate": 0.1,
+            "num_leaves": 15,
+            "verbose": -1
+        }
+    },
+    "order_types": {
+        "entry": "market",
+        "exit": "market",
+        "stoploss": "limit",
+        "stoploss_on_exchange": false
+    },
+    "order_time_in_force": {
+        "entry": "gtc",
+        "exit": "gtc"
+    },
+    "entry_pricing": {
+        "price_side": "other",
+        "use_order_book": false,
+        "price_last_balance": 0.0
+    },
+    "exit_pricing": {
+        "price_side": "other",
+        "use_order_book": false
+    },
+    "use_exit_signal": true,
+    "exit_profit_only": false,
+    "stoploss": -0.05,
+    "stake_currency": "USDT",
+    "stake_amount": 150,
+    "max_open_trades": 4,
+    "unfilledtimeout": {
+        "entry": 5,
+        "exit": 15,
+        "unit": "minutes"
+    },
+    "fee": 0.0008
+}

config_examples/config_binance.example.json_diabled

@@ -1,84 +0,0 @@
-{
-    "$schema": "https://schema.freqtrade.io/schema.json",
-    "max_open_trades": 3,
-    "stake_currency": "USDT",
-    "stake_amount": 0.05,
-    "tradable_balance_ratio": 0.99,
-    "fiat_display_currency": "USD",
-    "timeframe": "5m",
-    "dry_run": true,
-    "cancel_open_orders_on_exit": false,
-    "unfilledtimeout": {
-        "entry": 10,
-        "exit": 10,
-        "exit_timeout_count": 0,
-        "unit": "minutes"
-    },
-    "entry_pricing": {
-        "price_side": "same",
-        "use_order_book": true,
-        "order_book_top": 1,
-        "price_last_balance": 0.0,
-        "check_depth_of_market": {
-            "enabled": false,
-            "bids_to_ask_delta": 1
-        }
-    },
-    "exit_pricing": {
-        "price_side": "same",
-        "use_order_book": true,
-        "order_book_top": 1
-    },
-    "exchange": {
-        "name": "binance",
-        "key": "your_exchange_key",
-        "secret": "your_exchange_secret",
-        "ccxt_config": {},
-        "ccxt_async_config": {
-        },
-        "pair_whitelist": [
-            "ALGO/USDT",
-            "ATOM/USDT",
-            "BAT/USDT",
-            "BCH/USDT",
-            "BRD/USDT",
-            "EOS/USDT",
-            "ETH/USDT",
-            "IOTA/USDT",
-            "LINK/USDT",
-            "LTC/USDT",
-            "NEO/USDT",
-            "NXS/USDT",
-            "XMR/USDT",
-            "XRP/USDT",
-            "XTZ/USDT"
-        ],
-        "pair_blacklist": [
-            "BNB/.*"
-        ]
-    },
-    "pairlists": [
-        {"method": "StaticPairList"}
-    ],
-    "telegram": {
-        "enabled": false,
-        "token": "your_telegram_token",
-        "chat_id": "your_telegram_chat_id"
-    },
-    "api_server": {
-        "enabled": false,
-        "listen_ip_address": "127.0.0.1",
-        "listen_port": 8080,
-        "verbosity": "error",
-        "jwt_secret_key": "somethingrandom",
-        "CORS_origins": [],
-        "username": "freqtrader",
-        "password": "SuperSecurePassword"
-    },
-    "bot_name": "freqtrade",
-    "initial_state": "running",
-    "force_entry_enable": false,
-    "internals": {
-        "process_throttle_secs": 5
-    }
-}

config_examples/config_full.example.json_disabled

@@ -1,212 +0,0 @@
-{
-    "$schema": "https://schema.freqtrade.io/schema.json",
-    "max_open_trades": 3,
-    "stake_currency": "BTC",
-    "stake_amount": 0.05,
-    "tradable_balance_ratio": 0.99,
-    "fiat_display_currency": "USD",
-    "amount_reserve_percent": 0.05,
-    "available_capital": 1000,
-    "amend_last_stake_amount": false,
-    "last_stake_amount_min_ratio": 0.5,
-    "dry_run": true,
-    "dry_run_wallet": 1000,
-    "cancel_open_orders_on_exit": false,
-    "timeframe": "5m",
-    "trailing_stop": false,
-    "trailing_stop_positive": 0.005,
-    "trailing_stop_positive_offset": 0.0051,
-    "trailing_only_offset_is_reached": false,
-    "use_exit_signal": true,
-    "exit_profit_only": false,
-    "exit_profit_offset": 0.0,
-    "ignore_roi_if_entry_signal": false,
-    "ignore_buying_expired_candle_after": 300,
-    "trading_mode": "spot",
-    "margin_mode": "",
-    "minimal_roi": {
-        "40":  0.0,
-        "30":  0.01,
-        "20":  0.02,
-        "0":  0.04
-    },
-    "stoploss": -0.10,
-    "unfilledtimeout": {
-        "entry": 10,
-        "exit": 10,
-        "exit_timeout_count": 0,
-        "unit": "minutes"
-    },
-    "entry_pricing": {
-        "price_side": "same",
-        "use_order_book": true,
-        "order_book_top": 1,
-        "price_last_balance": 0.0,
-        "check_depth_of_market": {
-            "enabled": false,
-            "bids_to_ask_delta": 1
-        }
-    },
-    "exit_pricing":{
-        "price_side": "same",
-        "use_order_book": true,
-        "order_book_top": 1,
-        "price_last_balance": 0.0
-    },
-    "order_types": {
-        "entry": "limit",
-        "exit": "limit",
-        "emergency_exit": "market",
-        "force_exit": "market",
-        "force_entry": "market",
-        "stoploss": "market",
-        "stoploss_on_exchange": false,
-        "stoploss_price_type": "last",
-        "stoploss_on_exchange_interval": 60,
-        "stoploss_on_exchange_limit_ratio": 0.99
-    },
-    "order_time_in_force": {
-        "entry": "GTC",
-        "exit": "GTC"
-    },
-    "pairlists": [
-        {"method": "StaticPairList"},
-        {"method": "FullTradesFilter"},
-        {
-            "method": "VolumePairList",
-            "number_assets": 20,
-            "sort_key": "quoteVolume",
-            "refresh_period": 1800
-        },
-        {"method": "AgeFilter", "min_days_listed": 10},
-        {"method": "PrecisionFilter"},
-        {"method": "PriceFilter", "low_price_ratio": 0.01, "min_price": 0.00000010},
-        {"method": "SpreadFilter", "max_spread_ratio": 0.005},
-        {
-            "method": "RangeStabilityFilter",
-            "lookback_days": 10,
-            "min_rate_of_change": 0.01,
-            "refresh_period": 1440
-        }
-    ],
-    "exchange": {
-        "name": "binance",
-        "key": "your_exchange_key",
-        "secret": "your_exchange_secret",
-        "password": "",
-        "log_responses": false,
-        // "unknown_fee_rate": 1,
-        "ccxt_config": {},
-        "ccxt_async_config": {},
-        "pair_whitelist": [
-            "ALGO/BTC",
-            "ATOM/BTC",
-            "BAT/BTC",
-            "BCH/BTC",
-            "BRD/BTC",
-            "EOS/BTC",
-            "ETH/BTC",
-            "IOTA/BTC",
-            "LINK/BTC",
-            "LTC/BTC",
-            "NEO/BTC",
-            "NXS/BTC",
-            "XMR/BTC",
-            "XRP/BTC",
-            "XTZ/BTC"
-        ],
-        "pair_blacklist": [
-            "DOGE/BTC"
-        ],
-        "outdated_offset": 5,
-        "markets_refresh_interval": 60
-    },
-    "edge": {
-        "enabled": false,
-        "process_throttle_secs": 3600,
-        "calculate_since_number_of_days": 7,
-        "allowed_risk": 0.01,
-        "stoploss_range_min": -0.01,
-        "stoploss_range_max": -0.1,
-        "stoploss_range_step": -0.01,
-        "minimum_winrate": 0.60,
-        "minimum_expectancy": 0.20,
-        "min_trade_number": 10,
-        "max_trade_duration_minute": 1440,
-        "remove_pumps": false
-    },
-    "telegram": {
-        "enabled": false,
-        "token": "your_telegram_token",
-        "chat_id": "your_telegram_chat_id",
-        "notification_settings": {
-            "status": "on",
-            "warning": "on",
-            "startup": "on",
-            "entry": "on",
-            "entry_fill": "on",
-            "exit": {
-                "roi": "off",
-                "emergency_exit": "off",
-                "force_exit": "off",
-                "exit_signal": "off",
-                "trailing_stop_loss": "off",
-                "stop_loss": "off",
-                "stoploss_on_exchange": "off",
-                "custom_exit": "off"
-            },
-            "exit_fill": "on",
-            "entry_cancel": "on",
-            "exit_cancel": "on",
-            "protection_trigger": "off",
-            "protection_trigger_global": "on",
-            "show_candle": "off"
-        },
-        "reload": true,
-        "balance_dust_level": 0.01
-    },
-    "api_server": {
-        "enabled": false,
-        "listen_ip_address": "127.0.0.1",
-        "listen_port": 8080,
-        "verbosity": "error",
-        "enable_openapi": false,
-        "jwt_secret_key": "somethingrandom",
-        "CORS_origins": [],
-        "username": "freqtrader",
-        "password": "SuperSecurePassword",
-        "ws_token": "secret_ws_t0ken."
-    },
-    "external_message_consumer": {
-        "enabled": false,
-        "producers": [
-          {
-            "name": "default",
-            "host": "127.0.0.2",
-            "port": 8080,
-            "ws_token": "secret_ws_t0ken."
-          }
-        ],
-        "wait_timeout": 300,
-        "ping_timeout": 10,
-        "sleep_time": 10,
-        "remove_entry_exit_signals": false,
-        "message_size_limit": 8
-    },
-    "bot_name": "freqtrade",
-    "db_url": "sqlite:///tradesv3.sqlite",
-    "initial_state": "running",
-    "force_entry_enable": false,
-    "internals": {
-        "process_throttle_secs": 5,
-        "heartbeat_interval": 60
-    },
-    "disable_dataframe_checks": false,
-    "strategy": "SampleStrategy",
-    "strategy_path": "user_data/strategies/",
-    "recursive_strategy_search": false,
-    "add_config_files": [],
-    "reduce_df_footprint": false,
-    "dataformat_ohlcv": "feather",
-    "dataformat_trades": "feather"
-}

config_examples/config_kraken.example.json_disabled

@@ -1,90 +0,0 @@
-{
-    "$schema": "https://schema.freqtrade.io/schema.json",
-    "max_open_trades": 5,
-    "stake_currency": "EUR",
-    "stake_amount": 10,
-    "tradable_balance_ratio": 0.99,
-    "fiat_display_currency": "EUR",
-    "timeframe": "5m",
-    "dry_run": true,
-    "cancel_open_orders_on_exit": false,
-    "unfilledtimeout": {
-        "entry": 10,
-        "exit": 10,
-        "exit_timeout_count": 0,
-        "unit": "minutes"
-    },
-    "entry_pricing": {
-        "price_side": "same",
-        "use_order_book": true,
-        "order_book_top": 1,
-        "price_last_balance": 0.0,
-        "check_depth_of_market": {
-            "enabled": false,
-            "bids_to_ask_delta": 1
-        }
-    },
-    "exit_pricing":{
-        "price_side": "same",
-        "use_order_book": true,
-        "order_book_top": 1
-    },
-    "exchange": {
-        "name": "kraken",
-        "key": "your_exchange_key",
-        "secret": "your_exchange_key",
-        "ccxt_config": {},
-        "ccxt_async_config": {
-        },
-        "pair_whitelist": [
-            "ADA/EUR",
-            "ATOM/EUR",
-            "BAT/EUR",
-            "BCH/EUR",
-            "BTC/EUR",
-            "DAI/EUR",
-            "DASH/EUR",
-            "EOS/EUR",
-            "ETC/EUR",
-            "ETH/EUR",
-            "LINK/EUR",
-            "LTC/EUR",
-            "QTUM/EUR",
-            "REP/EUR",
-            "WAVES/EUR",
-            "XLM/EUR",
-            "XMR/EUR",
-            "XRP/EUR",
-            "XTZ/EUR",
-            "ZEC/EUR"
-        ],
-        "pair_blacklist": [
-
-        ]
-    },
-    "pairlists": [
-        {"method": "StaticPairList"}
-    ],
-    "telegram": {
-        "enabled": false,
-        "token": "your_telegram_token",
-        "chat_id": "your_telegram_chat_id"
-    },
-    "api_server": {
-        "enabled": false,
-        "listen_ip_address": "127.0.0.1",
-        "listen_port": 8080,
-        "verbosity": "error",
-        "jwt_secret_key": "somethingrandom",
-        "CORS_origins": [],
-        "username": "freqtrader",
-        "password": "SuperSecurePassword"
-    },
-    "bot_name": "freqtrade",
-    "initial_state": "running",
-    "force_entry_enable": false,
-    "internals": {
-        "process_throttle_secs": 5
-    },
-    "download_trades": true
-}

config_examples/freqaiprimer.json

@@ -0,0 +1,145 @@
+{
+    "$schema": "https://schema.freqtrade.io/schema.json",
+    "trading_mode": "spot",
+    "margin_mode": "isolated",
+    "max_open_trades": 10,
+    "stake_currency": "USDT",
+    "stake_amount": 55,
+    "tradable_balance_ratio": 1,
+    "fiat_display_currency": "USD",
+    "dry_run": true,
+    "timeframe": "3m",
+    "dry_run_wallet": 2000,
+    "cancel_open_orders_on_exit": true,
+    "stoploss": -0.14,
+    "max_entry_position_adjustment": 4,
+    "position_adjustment_enable": true,
+    "amount_reserve_percent": 0.05,
+    "unfilledtimeout": {
+        "entry": 5,
+        "exit": 15
+    },
+    "dry_run": false,
+    "startup_candle_count": 200,
+    "tradable_balance_ratio": 0.5,
+    "exchange": {
+        "name": "okx",
+        "key": "314e75b0-1113-47e8-ad01-1fca7e3c0496",
+        "secret": "9C8B170390F46EA6FB87592AD46F5A34",
+	      "password": "nekFoylf:Om0",
+        "enable_ws": false,
+        "ccxt_config": {
+            "enableRateLimit": true,
+            "rateLimit": 500,
+            "options": {
+                "defaultType": "spot"
+            }
+        },
+        "ccxt_async_config": {
+            "enableRateLimit": true,
+            "rateLimit": 3000,
+            "timeout": 20000
+        },
+        "pair_whitelist": ["BTC/USDT",  "TON/USDT", "DOT/USDT", "XRP/USDT",  "OKB/USDT", "SOL/USDT", "DOGE/USDT", "WCT/USDT", "TRUMP/USDT", "SUI/USDT", "PEPE/USDT", "TRB/USDT", "MASK/USDT", "UNI/USDT", "KAITO/USDT"],
+        "pair_blacklist": []
+    },
+    "entry_pricing": {
+        "price_side": "same",
+        "use_order_book": true,
+        "order_book_top": 1,
+        "price_last_balance": 0.0,
+        "check_depth_of_market": {
+            "enabled": false,
+            "bids_to_ask_delta": 1
+        }
+    },
+    "exit_pricing": {
+        "price_side": "other",
+        "use_order_book": true,
+        "order_book_top": 1
+    },
+    "pairlists": [
+        {
+            "method": "StaticPairList"
+        }    
+    ],
+    "freqai": {
+        "enabled": true,
+        "data_kitchen": {
+            "fillna": "ffill"
+        },
+        "freqaimodel": "LightGBMMultiTargetRegressor",
+        "purge_old_models": 2,
+        "identifier": "test58",
+        "train_period_days": 7,
+        "backtest_period_days": 1,
+        "live_retrain_hours": 2,
+        "outlier_detection": {
+            "method": "IsolationForest",
+            "contamination": 0.1
+        },
+        "feature_selection": {
+          "method": "recursive_elimination"
+        },
+        "feature_parameters": {
+            "include_timeframes": [
+                "3m",
+                "15m",
+                "1h",
+            ],
+            "include_corr_pairlist": [
+                "BTC/USDT",
+                "SOL/USDT"
+            ],
+            "outlier_protection_percentage": 0.1,
+            "label_period_candles": 24,
+            "include_shifted_candles": 3,
+            "DI_threshold": 0.7,
+            "weight_factor": 0.9,
+            "principal_component_analysis": false,
+            "use_SVM_to_remove_outliers": true,
+            "indicator_periods_candles": [
+                10,
+                20,
+                50
+            ],
+            "outlier_protection_percentage": 0.15,
+            "plot_feature_importances": 10
+        },
+        "data_split_parameters": {
+            "test_size": 0.2,
+              "shuffle": false
+        },
+        "model_training_parameters": {
+            "n_estimators": 400,
+            "learning_rate": 0.03,
+            "num_leaves": 40,
+            "max_depth": 10,
+            "min_data_in_leaf": 20,
+            "feature_fraction": 0.8,
+            "bagging_fraction": 0.8,
+            "bagging_freq": 5,
+            "verbose": -1
+        }
+    },
+    "api_server": {
+        "enabled": true,
+        "listen_ip_address": "0.0.0.0",
+        "listen_port": 8080,
+        "verbosity": "error",
+        "enable_openapi": false,
+        "jwt_secret_key": "6a599ab046dbb419014807dffd7b8823bfa7e5df56b17d545485deb87331b4ca",
+        "ws_token": "6O5pBDiRigiZrmIsofaE2rkKMJtf9h8zVQ",
+        "CORS_origins": [],
+        "username": "freqAdmin",
+        "password": "admin"
+    },
+    "bot_name": "freqtrade",
+    "initial_state": "running",
+    "force_entry_enable": false,
+    "internals": {
+        "process_throttle_secs": 5,
+        "heartbeat_interval": 20,
+        "loglevel": "DEBUG"
+    }
+}

config_examples/live.json

@@ -0,0 +1,48 @@
+{
+      "exchange": {
+        "name": "okx",
+        "key": "cbda9fde-b9e3-4a2d-94f9-e5c3705dfb5c",
+        "secret": "CD3B7DB459CBBD540E0926E5C48150E1",
+	      "password": "c^-d:g;P2S9?Q#^",
+        "enable_ws": false,
+        "ccxt_config": {
+            "enableRateLimit": true,
+            "rateLimit": 500,
+            "options": {
+                "defaultType": "spot"
+            }
+        },
+        "ccxt_async_config": {
+            "enableRateLimit": true,
+            "rateLimit": 3000,
+            "timeout": 20000
+        },
+        "pair_whitelist": ["BTC/USDT",  "TON/USDT", "DOT/USDT", "XRP/USDT",  "OKB/USDT", "SOL/USDT", "DOGE/USDT", "WCT/USDT", "TRUMP/USDT", "SUI/USDT", "PEPE/USDT", "TRB/USDT", "MASK/USDT", "UNI/USDT", "KAITO/USDT"],
+        "pair_blacklist": []
+    },
+    "pairlists": [
+        {
+            "method": "StaticPairList"
+        }
+    ],
+    "order_types": {
+        "entry": "market",
+        "exit": "market",
+        "stoploss": "limit",
+        "stoploss_on_exchange": false
+    },
+    "order_time_in_force": {
+        "entry": "gtc",
+        "exit": "gtc"
+    },
+    "entry_pricing": {
+        "price_side": "other",
+        "use_order_book": false,
+        "price_last_balance": 0.0
+    },
+    "exit_pricing": {
+        "price_side": "other",
+        "use_order_book": false
+    },
+    "fee": 0.0008
+}

config_examples/theforcev7.json

@@ -0,0 +1,96 @@
+{
+    "$schema": "https://schema.freqtrade.io/schema.json",
+    "trading_mode": "spot",
+    "margin_mode": "isolated",
+    "max_open_trades": 3,
+    "stake_amount": 300,
+    "unfilledtimeout": {
+        "entry": 10,
+        "exit": 20
+    },
+    "entry_pricing": {
+        "price_side": "same",
+        "use_order_book": true,
+        "order_book_top": 1
+    },
+    "exit_pricing": {
+        "price_side": "other",
+        "use_order_book": true,
+        "order_book_top": 1
+    },
+    "stake_currency": "USDT",
+    "trailing_stop": false,
+    "use_custom_stoploss": true,
+    "startup_candle_count": 200,
+    "tradable_balance_ratio": 1,
+    "fiat_display_currency": "USD",
+    "dry_run": true,
+    "identifier": "demo1",
+    "timeframe": "5m",
+    "dry_run_wallet": 1000,
+    "cancel_open_orders_on_exit": true,
+    "stoploss": -0.061,
+    "roi": {},
+    "trailing": {
+      "trailing_stop": false,
+      "trailing_stop_positive": null,
+      "trailing_stop_positive_offset": 0.0,
+      "trailing_only_offset_is_reached": false
+    },
+    "max_open_trades": 3,
+    "buy": {
+      "adx_buy": 25,
+      "atr_ratio": 0.005
+    },
+    "sell": {
+      "ema_fast_period": 7,
+      "rsi_sell": 60
+    },
+    "exchange": {
+        "name": "okx",
+        "key": "REDACTED",
+        "secret": "REDACTED",
+        "enable_ws": false,
+        "ccxt_config": {
+            "enableRateLimit": true,
+            "rateLimit": 500,
+            "options": {
+                "defaultType": "spot"
+            }
+        },
+        "ccxt_async_config": {
+            "enableRateLimit": true,
+            "rateLimit": 3000,
+            "timeout": 20000
+        },
+        "pair_whitelist": ["BTC/USDT", "ETH/USDT", "SOL/USDT", "DOT/USDT"],
+        "pair_blacklist": []
+    },
+    "pairlists": [
+        {
+            "method": "StaticPairList"
+        }
+    ],
+    "api_server": {
+        "enabled": true,
+        "listen_ip_address": "0.0.0.0",
+        "listen_port": 8080,
+        "verbosity": "error",
+        "enable_openapi": false,
+        "jwt_secret_key": "6a599ab046dbb419014807dffd7b8823bfa7e5df56b17d545485deb87331b4ca",
+        "ws_token": "6O5pBDiRigiZrmIsofaE2rkKMJtf9h8zVQ",
+        "CORS_origins": [],
+        "username": "freqAdmin",
+        "password": "REDACTED"
+    },
+    "use_exit_signal": true,
+    "bot_name": "freqtrade",
+    "initial_state": "running",
+    "force_entry_enable": false,
+    "internals": {
+        "process_throttle_secs": 5,
+        "heartbeat_interval": 20,
+        "loglevel": "INFO"
+    },
+    "config_files": ["config_examples/theforcev7.json"]
+}

doc/README.md

@@ -0,0 +1,128 @@
+### 策略概述：FreqaiPrimer
+
+**FreqaiPrimer** 是一个基于 **FreqAI**（频率人工智能）框架的交易策略，结合机器学习预测与传统技术指标，旨在适应多种市场环境（震荡、下跌、单边上涨）。策略以 3 分钟（3m）为主时间框架，整合 3m、15m 和 1h 数据，通过动态调整入场、退出条件和下注金额，优化交易机会并控制风险。核心目标是：
+1. **捕捉价值背离**：利用 FreqAI 预测价格与长期均线（EMA200）的偏离（`&-price_value_divergence`），识别超买/超卖点。
+2. **适应趋势市场**：通过趋势得分（`trend_score`，0-100，100 为强牛市）动态调整买卖条件，支持单边上涨行情。
+3. **风险管理**：通过动态仓位调整、追踪止损和冷却期，平衡收益与风险。
+4. **多指标融合**：结合 RSI、布林带、MACD、Stochastic 和 ADX 等指标，增强信号可靠性。
+
+策略使用 LightGBM 回归模型预测价值背离，结合技术指标（如 EMA、RSI、MACD、Stochastic、ADX）生成买卖信号，并通过 Hyperopt 优化参数（如阈值、止损、仓位调整）。以下是入场和退出逻辑的详细描述。
+
+---
+
+### 入场逻辑
+入场逻辑在 `populate_entry_trend` 方法中实现，分为 **震荡/下跌市场** 和 **趋势市场** 两种条件，通过趋势得分（`trend_score`）动态调整，适应不同市场环境。入场信号通过 `enter_long = 1` 标记。
+
+#### 1. 震荡/下跌市场入场条件
+适用于震荡或下跌市场，旨在捕捉超卖机会：
+- **价值背离（`&-price_value_divergence`）**：
+  - 由 FreqAI 预测，衡量价格与 EMA200 的偏离百分比。
+  - 条件：`&-price_value_divergence < dynamic_buy_threshold`。
+  - 阈值：根据 `trend_score` 在 -0.1（熊市）到 0.02（牛市）间动态调整，牛市允许正背离。
+  - 含义：价格显著低于 EMA200，表明超卖。
+- **成交量标准化得分（`volume_z_score`）**：
+  - 计算成交量相对于 20 周期均值的标准化得分。
+  - 条件：`volume_z_score > volume_z_score_threshold`（0.5 到 2.0，牛市阈值较低）。
+  - 含义：高成交量确认市场活跃，可能伴随反转。
+- **RSI**：
+  - 使用 14 周期 RSI。
+  - 条件：`rsi < rsi_threshold`（40 到 60，牛市阈值较低）。
+  - 含义：低 RSI 表明超卖状态。
+- **布林带下轨**：
+  - 条件：`close <= bb_lowerband`（20 周期布林带，2 倍标准差）。
+  - 含义：价格触及下轨，强化超卖信号。
+
+**组合逻辑**：所有条件需同时满足（`cond1 & cond2 & cond3 & cond4`）。
+
+#### 2. 趋势市场入场条件（`trend_score > 70`）
+为捕捉单边上涨行情中的回调买入机会，新增以下条件：
+- **价格接近 EMA20**：
+  - 条件：`close >= ema20 * 0.995`（允许 0.5% 回调）。
+  - 含义：价格在短期均线附近，适合趋势回调买入。
+- **EMA20 上穿 EMA50**：
+  - 条件：`ema20 > ema50`。
+  - 含义：短期趋势强于中期趋势，确认上涨趋势。
+- **MACD 看涨交叉**：
+  - 条件：`macd > macdsignal` 且前一根 K 线 `macd <= macdsignal`。
+  - 含义：短期动能增强，适合入场。
+- **Stochastic 慢线回升**：
+  - 条件：`slowk > slowk.shift(1)` 且 `slowk < 70`。
+  - 含义：慢线上升且未超买，表明回调结束。
+
+**组合逻辑**：所有趋势条件需同时满足（`cond_trend1 & cond_trend2 & cond_trend3 & cond_trend4`），仅在强牛市（`trend_score > 70`）启用。
+
+#### 3. 合并逻辑
+- 震荡和趋势条件通过 OR 逻辑合并（`oscillation_condition | trend_condition`），允许任一条件触发入场。
+- 动态冷却期（1-10 分钟，牛市缩短，熊市延长）通过 `COOLDOWN_PERIOD_MINUTES` 控制连续入场间隔。
+- 下注金额通过 `custom_stake_amount` 根据 `trend_score` 动态调整（1-200 倍映射，牛市增加，熊市减少）。
+
+---
+
+### 退出逻辑
+退出逻辑在 `populate_exit_trend` 和 `adjust_trade_position` 方法中实现，分为 **震荡/下跌市场** 和 **趋势市场** 条件，结合 ADX 确认趋势反转，优化单边上涨行情中的持有时间。退出信号通过 `exit_long = 1` 或仓位调整触发。
+
+#### 1. 震荡/下跌市场退出条件
+适用于震荡或下跌市场，捕捉超买或获利点：
+- **价值背离（`&-price_value_divergence`）**：
+  - 条件：`&-price_value_divergence > dynamic_sell_threshold`。
+  - 阈值：根据 `trend_score` 在 0.001（熊市）到 0.1（牛市）间动态调整，牛市允许更大正背离。
+  - 含义：价格显著高于 EMA200，表明超买。
+- **RSI**：
+  - 条件：`rsi > rsi_threshold`（75 到 85，牛市阈值更高）。
+  - 含义：高 RSI 表明超买状态。
+
+**组合逻辑**：任一条件满足（`cond1 | cond2`）触发卖出。
+
+#### 2. 趋势市场退出条件（`trend_score > 70`）
+为避免牛市中过早退出，加入趋势反转确认：
+- **MACD 看跌交叉**：
+  - 条件：`macd < macdsignal` 且前一根 K 线 `macd >= macdsignal`。
+  - 含义：短期动能减弱，可能反转。
+- **Stochastic 超买回落**：
+  - 条件：`slowk > 80` 且 `slowk < slowk.shift(1)`。
+  - 含义：慢线超买后下降，表明趋势可能结束。
+- **EMA20 下穿 EMA50**：
+  - 条件：`ema20 < ema50` 且前一根 K 线 `ema20 >= ema50`。
+  - 含义：短期趋势弱于中期趋势，确认反转。
+- **ADX 趋势减弱**：
+  - 条件：`adx > 25` 且 `adx_diff < 0`（ADX 下降）。
+  - 含义：趋势强度减弱，可能预示反转。
+
+**组合逻辑**：任一趋势条件满足（`cond_trend1 | cond_trend2 | cond_trend3 | cond_trend4`）触发卖出，仅在强牛市启用。
+
+#### 3. 合并逻辑与仓位调整
+- 震荡和趋势条件通过 OR 逻辑合并（`oscillation_condition | trend_condition`）。
+- **分级止盈**（`adjust_trade_position`）：
+  - 利润 ≥ 3%：减仓（比例 0.6-1.0，牛市较低）。
+  - 利润 ≥ 7%（牛市）：进一步减仓（比例 0.8-1.2），保留仓位捕捉更大上涨。
+- **追踪止损**：
+  - 启动点（`TRAILING_STOP_START`）：牛市提高（如 3% → 4.5%）。
+  - 止损距离（`TRAILING_STOP_DISTANCE`）：牛市增大（如 1% → 2%）。
+  - 动态调整：根据 `trend_score` 使用 Sigmoid 函数，牛市更宽松。
+- **确认退出**（`confirm_trade_exit`）：记录退出原因，调整卖出价格（+0.125% 优化收益）。
+
+---
+
+### 策略特点
+1. **多时间框架**：整合 3m、15m、1h 数据，增强特征工程和趋势判断。
+2. **动态适应**：趋势得分动态调整阈值、下注金额和冷却期，牛市宽松，熊市严格。
+3. **多指标融合**：结合 FreqAI 预测（价值背离）、RSI、布林带、MACD、Stochastic 和 ADX。
+4. **风险管理**：
+   - 固定止损：-1.5%。
+   - 动态仓位：支持加仓（最多 3 次，倍数 2、4、8）和减仓。
+   - 追踪止损：牛市中延迟触发，容忍更大回调。
+5. **调试支持**：详细日志记录条件状态，便于分析。
+
+---
+
+### 单边上涨行情优化
+- **入场**：新增趋势条件（EMA20、MACD、Stochastic），允许在牛市回调点（如价格接近 EMA20）入场，解决震荡条件过于严格的问题。
+- **退出**：动态放宽 RSI（75 → 85）和背离阈值（0.05 → 0.1），加入 ADX（趋势减弱）、MACD、Stochastic 和均线反转信号，延迟牛市退出，捕捉更大利润。
+- **协调性**：入场和退出逻辑均基于 `trend_score` 和多指标确认，确保牛市中延长持有时间并及时止盈。
+
+---
+
+### 总结
+**FreqaiPrimer** 通过 FreqAI 预测价值背离，结合 RSI、布林带、MACD、Stochastic 和 ADX 等指标，动态适应震荡和趋势市场。入场逻辑在震荡市场捕捉超卖机会，在牛市中利用趋势指标捕捉回调买入点；退出逻辑通过动态阈值和趋势反转信号（包括 ADX 确认），避免牛市过早退出，同时在趋势减弱时及时止盈。策略通过动态仓位、追踪止损和冷却期优化风险管理，适合加密货币等高波动市场。
+
+如需进一步细化逻辑、提供回测代码或分析，请告诉我！

doc/hyperopts.md

@@ -0,0 +1,14 @@
+## hyperopts如何使用
+
+每周运行一次 hyperopt, 每次运行最近2个月的数据, 得到 stoploss 最优解后 手动更新 config, 提交代码
+
+假设今天是2025年5月15日
+```
+cd tools
+./download.sh 
+./hyperopt.sh 20250314 20250516
+然后得到结果, stoploss 为 -0.06 写到log里
+```
+
+后面改成自动更新并不麻烦
+策略目录有自动生成的json, 是hypertop计算结果, 需脱掉一层 paramas后 方可使用

doc/mark.md

@@ -0,0 +1,58 @@
+### 优化记录
+
+
+- 20250606 10:52
+
+当前hyperopt使用 ./hyperopt 20250425 20250525 得到的优化结果, 然后
+
+./backtest.sh 20250520 20250525, 结果:
+
+                   SUMMARY METRICS
+┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━┓
+┃ Metric                      ┃ Value               ┃
+┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━┩
+│ Backtesting from            │ 2025-05-20 00:00:00 │
+│ Backtesting to              │ 2025-05-25 00:00:00 │
+│ Trading Mode                │ Spot                │
+│ Max open trades             │ 4                   │
+│                             │                     │
+│ Total/Daily Avg Trades      │ 16 / 3.2            │
+│ Starting balance            │ 1000 USDT           │
+│ Final balance               │ 1010.711 USDT       │
+│ Absolute profit             │ 10.711 USDT         │
+│ Total profit %              │ 1.07%               │
+│ CAGR %                      │ 117.65%             │
+│ Sortino                     │ 53.06               │
+│ Sharpe                      │ 11.07               │
+│ Calmar                      │ 349.31              │
+│ SQN                         │ 0.70                │
+│ Profit factor               │ 1.61                │
+│ Expectancy (Ratio)          │ 0.67 (-0.07)        │
+│ Avg. daily profit %         │ 0.21%               │
+│ Avg. stake amount           │ 149.999 USDT        │
+│ Total trade volume          │ 4818.39 USDT        │
+│                             │                     │
+│ Best Pair                   │ TRUMP/USDT 0.98%    │
+│ Worst Pair                  │ MASK/USDT -0.85%    │
+│ Best trade                  │ TRUMP/USDT 5.31%    │
+│ Worst trade                 │ UNI/USDT -4.60%     │
+│ Best day                    │ 15.174 USDT         │
+│ Worst day                   │ -6.004 USDT         │
+│ Days win/draw/lose          │ 2 / 0 / 2           │
+│ Avg. Duration Winners       │ 1:08:00             │
+│ Avg. Duration Loser         │ 0:18:00             │
+│ Max Consecutive Wins / Loss │ 3 / 3               │
+│ Rejected Entry signals      │ 0                   │
+│ Entry/Exit Timeouts         │ 9 / 0               │
+│                             │                     │
+│ Min balance                 │ 1001.8 USDT         │
+│ Max balance                 │ 1020.574 USDT       │
+│ Max % of account underwater │ 1.17%               │
+│ Absolute Drawdown (Account) │ 1.17%               │
+│ Absolute Drawdown           │ 11.957 USDT         │
+│ Drawdown high               │ 20.574 USDT         │
+│ Drawdown low                │ 8.617 USDT          │
+│ Drawdown Start              │ 2025-05-22 20:48:00 │
+│ Drawdown End                │ 2025-05-24 17:27:00 │
+│ Market change               │ 0.81%               │
+└─────────────────────────────┴─────────────────────┘

docker-compose.yml

@@ -1,7 +1,7 @@
----
+version: '3.8'
 services:
   freqtrade:
-    image: freqtradeorg/freqtrade:stable_freqaitorch
+    image: freqtradeorg/freqtrade:develop_freqaitorch
     # # Enable GPU Image and GPU Resources
     # # Make sure to uncomment the whole deploy section
     # deploy:
@@ -23,30 +23,31 @@ services:
       - "./config_examples:/freqtrade/config_examples"
       - "./freqtrade/templates:/freqtrade/templates"
       - "./freqtrade/exchange/:/freqtrade/exchange"
+      - "./ccxt/async_support/okx.py:/home/ftuser/.local/lib/python3.12/site-packages/ccxt/async_support/okx.py"
     # Expose api on port 8080 (localhost only)
     # Please read the https://www.freqtrade.io/en/stable/rest-api/ documentation
     # for more information.
     ports:
       - "8080:8080"
     # Default command used when running `docker compose up`
-      
+
     # --freqaimodel XGBoostRegressor
-    # commangd: >
-    #      # trade
-      # --logfile /freqtrade/user_data/logs/freqtrade.log
-      # --db-url sqlite:////freqtrade/user_data/tradesv3.sqlite
-      # --freqaimodel LightGBMRegressor
-      # --config /freqtrade/config_examples/config_freqai.okx.json
-      # --strategy FreqaiExampleStrategy
-      #  --strategy FreqaiExampleHybridStrategy
-      # --strategy-path /freqtrade/templates
+    # command: >
+    #   trade
+    #     --logfile /freqtrade/user_data/logs/freqtrade.log
+    #     --db-url sqlite:////freqtrade/user_data/tradesv3.sqlite
+    #     --freqaimodel LightGBMRegressor
+    #     --config /freqtrade/config_examples/config_freqai.okx.json
+    #     --strategy FreqaiExampleStrategy
+    #     --strategy-path /freqtrade/templates
     # command: >
     #   backtesting
     #     --logfile /freqtrade/user_data/logs/freqtrade.log
-    #     --freqaimodel XGBoostRegressor
+    #     --freqaimodel LightGBMRegressor
     #     --config /freqtrade/config_examples/config_freqai.okx.json
     #     --strategy-path /freqtrade/templates
     #     --strategy FreqaiExampleStrategy
+    #     --breakdown week month
     #     --timerange 20250310-20250410
     #     --export trades
     # command: >
@@ -58,15 +59,41 @@ services:
     #     --strategy FreqaiExampleStrategy
     #     --timerange 20250301-20250420
     #     --hyperopt-loss SharpeHyperOptLoss
-    #     --spaces roi stoploss 
+    #     --spaces roi stoploss trailing
     #     -e 200
 
-    command: >
-       backtesting
-        --logfile /freqtrade/user_data/logs/freqtrade.log
-        --freqaimodel LightGBMRegressor
-        --config /freqtrade/config_examples/config_freqai.okx.json
-        --config /freqtrade/templates/FreqaiExampleStrategy.json
-        --strategy-path /freqtrade/templates
-        --strategy FreqaiExampleStrategy
-        --timerange 20240920-20250420
+    # command: >
+    #   hyperopt
+    #     --logfile /freqtrade/user_data/logs/freqtrade.log
+    #     --freqaimodel LightGBMRegressor
+    #     --config /freqtrade/config_examples/config_freqai.okx.json
+    #     --strategy-path /freqtrade/templates
+    #     --strategy FreqaiExampleStrategy
+    #     --timerange 20250301-20250420
+    #     --hyperopt-loss SharpeHyperOptLoss
+    #     --spaces roi stoploss
+    #     -e 200
+
+    #--config /freqtrade/templates/FreqaiExampleStrategy.json
+
+    # command: >
+    #    backtesting
+    #     --logfile /freqtrade/user_data/logs/freqtrade.log
+    #     --freqaimodel XGBoostRegressor
+    #     --config /freqtrade/config_examples/config_my_hyperopt.json
+    #     --strategy-path /freqtrade/templates
+    #     --strategy TheForceV7
+    #     --timerange 20250301-20250405
+    #     --export trades
+    #     --fee 0.0008
+    #     --cache none
+    #
+    # command: >
+    #    trade
+    #     --logfile /freqtrade/user_data/logs/freqtrade.log
+    #     --db-url sqlite:////freqtrade/user_data/tradesv3.sqlite
+    #     --freqaimodel XGBoostRegressor
+    #     --config /freqtrade/config_examples/theforcev7.json
+    #     --strategy-path /freqtrade/templates
+    #     --strategy TheForceV7
+    #     --fee 0.0008

file-copy.sh

@@ -0,0 +1,47 @@
+import os
+import shutil
+
+# 定义源目录
+source_dir = "./result"
+
+# 定义目标目录和文件名
+target_strategy_path = "./freqtrade/template/FreqaiExampleStrategy.py"
+target_config_path = "./config_examples/config_freqai.okx.json"
+
+# 查找 *FreqaiExampleStrategy.py 文件
+strategy_file = None
+for file in os.listdir(source_dir):
+    if file.endswith("FreqaiExampleStrategy.py"):
+        strategy_file = os.path.join(source_dir, file)
+        break
+
+# 查找 *config.json 文件
+config_file = None
+for file in os.listdir(source_dir):
+    if file.endswith("config.json"):
+        config_file = os.path.join(source_dir, file)
+        break
+
+# 复制 *FreqaiExampleStrategy.py 文件
+if strategy_file:
+    try:
+        target_strategy_dir = os.path.dirname(target_strategy_path)
+        os.makedirs(target_strategy_dir, exist_ok=True)
+        shutil.copy2(strategy_file, target_strategy_path)
+        print(f"成功复制 {strategy_file} 到 {target_strategy_path}")
+    except Exception as e:
+        print(f"复制 {strategy_file} 时出错: {e}")
+else:
+    print("未找到 *FreqaiExampleStrategy.py 文件")
+
+# 复制 *config.json 文件
+if config_file:
+    try:
+        target_config_dir = os.path.dirname(target_config_path)
+        os.makedirs(target_config_dir, exist_ok=True)
+        shutil.copy2(config_file, target_config_path)
+        print(f"成功复制 {config_file} 到 {target_config_path}")
+    except Exception as e:
+        print(f"复制 {config_file} 时出错: {e}")
+else:
+    print("未找到 *config.json 文件")    

filter.py

@@ -0,0 +1,7 @@
+with open('output.log', 'r') as input_file, open('output_filted.log', 'w') as output_file:
+    for line in input_file:
+        if "validation_0" in line:
+            if "[99]" in line:
+                output_file.write(line)
+        else:
+            output_file.write(line)

filtered_output.log

@@ -0,0 +1,888 @@
+[99]	validation_0-rmse:0.13480	validation_1-rmse:0.09950
+[99]	validation_0-rmse:0.14079	validation_1-rmse:0.10036
+[99]	validation_0-rmse:0.15004	validation_1-rmse:0.09815
+[99]	validation_0-rmse:0.13402	validation_1-rmse:0.10289
+[99]	validation_0-rmse:0.14818	validation_1-rmse:0.10475
+[99]	validation_0-rmse:0.15738	validation_1-rmse:0.10531
+[99]	validation_0-rmse:0.14246	validation_1-rmse:0.11071
+[99]	validation_0-rmse:0.16321	validation_1-rmse:0.10649
+[99]	validation_0-rmse:0.15096	validation_1-rmse:0.10882
+[99]	validation_0-rmse:0.15694	validation_1-rmse:0.11396
+[99]	validation_0-rmse:0.18198	validation_1-rmse:0.10994
+[99]	validation_0-rmse:0.16824	validation_1-rmse:0.11403
+[99]	validation_0-rmse:0.16437	validation_1-rmse:0.11096
+[99]	validation_0-rmse:0.17746	validation_1-rmse:0.10819
+[99]	validation_0-rmse:0.15331	validation_1-rmse:0.11727
+[99]	validation_0-rmse:0.16802	validation_1-rmse:0.11547
+[99]	validation_0-rmse:0.15958	validation_1-rmse:0.11029
+[99]	validation_0-rmse:0.18742	validation_1-rmse:0.11149
+[99]	validation_0-rmse:0.15075	validation_1-rmse:0.09770
+[99]	validation_0-rmse:0.13818	validation_1-rmse:0.10038
+[99]	validation_0-rmse:0.12954	validation_1-rmse:0.09825
+[99]	validation_0-rmse:0.14837	validation_1-rmse:0.10398
+Creating freqtrade_freqtrade_run ... 
+Creating freqtrade_freqtrade_run ... done
+2025-04-29 01:54:55,246 - freqtrade - INFO - freqtrade 2025.3
+2025-04-29 01:54:55,464 - numexpr.utils - INFO - NumExpr defaulting to 12 threads.
+2025-04-29 01:54:56,878 - freqtrade.configuration.load_config - INFO - Using config: /freqtrade/config_examples/config_freqai.okx.json ...
+2025-04-29 01:54:56,879 - freqtrade.configuration.load_config - INFO - Using config: /freqtrade/templates/FreqaiExampleStrategy.json ...
+2025-04-29 01:54:56,881 - freqtrade.loggers - INFO - Enabling colorized output.
+2025-04-29 01:54:56,881 - root - INFO - Logfile configured
+2025-04-29 01:54:56,882 - freqtrade.loggers - INFO - Verbosity set to 0
+2025-04-29 01:54:56,882 - freqtrade.configuration.configuration - INFO - Using additional Strategy lookup path: /freqtrade/templates
+2025-04-29 01:54:56,883 - freqtrade.configuration.configuration - INFO - Using max_open_trades: 4 ...
+2025-04-29 01:54:56,883 - freqtrade.configuration.configuration - INFO - Parameter --timerange detected: 20250101-20250420 ...
+2025-04-29 01:54:56,907 - freqtrade.configuration.configuration - INFO - Using user-data directory: /freqtrade/user_data ...
+2025-04-29 01:54:56,908 - freqtrade.configuration.configuration - INFO - Using data directory: /freqtrade/user_data/data/okx ...
+2025-04-29 01:54:56,908 - freqtrade.configuration.configuration - INFO - Parameter --cache=none detected ...
+2025-04-29 01:54:56,908 - freqtrade.configuration.configuration - INFO - Filter trades by timerange: 20250101-20250420
+2025-04-29 01:54:56,909 - freqtrade.configuration.configuration - INFO - Using freqaimodel class name: XGBoostRegressor
+2025-04-29 01:54:56,910 - freqtrade.exchange.check_exchange - INFO - Checking exchange...
+2025-04-29 01:54:56,916 - freqtrade.exchange.check_exchange - INFO - Exchange "okx" is officially supported by the Freqtrade development team.
+2025-04-29 01:54:56,916 - freqtrade.configuration.configuration - INFO - Using pairlist from configuration.
+2025-04-29 01:54:56,917 - freqtrade.configuration.config_validation - INFO - Validating configuration ...
+2025-04-29 01:54:56,919 - freqtrade.commands.optimize_commands - INFO - Starting freqtrade in Backtesting mode
+2025-04-29 01:54:56,919 - freqtrade.exchange.exchange - INFO - Instance is running with dry_run enabled
+2025-04-29 01:54:56,920 - freqtrade.exchange.exchange - INFO - Using CCXT 4.4.69
+2025-04-29 01:54:56,920 - freqtrade.exchange.exchange - INFO - Applying additional ccxt config: {'enableRateLimit': True, 'rateLimit': 500, 'options': {'defaultType': 'spot'}}
+2025-04-29 01:54:56,925 - freqtrade.exchange.exchange - INFO - Applying additional ccxt config: {'enableRateLimit': True, 'rateLimit': 500, 'options': {'defaultType': 'spot'}, 'timeout': 20000}
+2025-04-29 01:54:56,931 - freqtrade.exchange.exchange - INFO - Using Exchange "OKX"
+2025-04-29 01:54:59,471 - freqtrade.resolvers.exchange_resolver - INFO - Using resolved exchange 'Okx'...
+2025-04-29 01:54:59,491 - freqtrade.resolvers.iresolver - INFO - Using resolved strategy FreqaiExampleStrategy from '/freqtrade/templates/FreqaiExampleStrategy.py'...
+2025-04-29 01:54:59,491 - freqtrade.strategy.hyper - INFO - Loading parameters from file /freqtrade/templates/FreqaiExampleStrategy.json
+2025-04-29 01:54:59,492 - freqtrade.resolvers.strategy_resolver - INFO - Override strategy 'timeframe' with value in config file: 3m.
+2025-04-29 01:54:59,492 - freqtrade.resolvers.strategy_resolver - INFO - Override strategy 'stoploss' with value in config file: -0.05.
+2025-04-29 01:54:59,493 - freqtrade.resolvers.strategy_resolver - INFO - Override strategy 'stake_currency' with value in config file: USDT.
+2025-04-29 01:54:59,493 - freqtrade.resolvers.strategy_resolver - INFO - Override strategy 'stake_amount' with value in config file: 150.
+2025-04-29 01:54:59,493 - freqtrade.resolvers.strategy_resolver - INFO - Override strategy 'startup_candle_count' with value in config file: 30.
+2025-04-29 01:54:59,494 - freqtrade.resolvers.strategy_resolver - INFO - Override strategy 'unfilledtimeout' with value in config file: {'entry': 5, 'exit': 15, 'exit_timeout_count': 0, 'unit': 
+'minutes'}.
+2025-04-29 01:54:59,494 - freqtrade.resolvers.strategy_resolver - INFO - Override strategy 'max_open_trades' with value in config file: 4.
+2025-04-29 01:54:59,494 - freqtrade.resolvers.strategy_resolver - INFO - Strategy using minimal_roi: {'0': 0.132, '8': 0.047, '14': 0.007, '60': 0}
+2025-04-29 01:54:59,495 - freqtrade.resolvers.strategy_resolver - INFO - Strategy using timeframe: 3m
+2025-04-29 01:54:59,495 - freqtrade.resolvers.strategy_resolver - INFO - Strategy using stoploss: -0.05
+2025-04-29 01:54:59,495 - freqtrade.resolvers.strategy_resolver - INFO - Strategy using trailing_stop: True
+2025-04-29 01:54:59,495 - freqtrade.resolvers.strategy_resolver - INFO - Strategy using trailing_stop_positive: 0.01
+2025-04-29 01:54:59,496 - freqtrade.resolvers.strategy_resolver - INFO - Strategy using trailing_stop_positive_offset: 0.02
+2025-04-29 01:54:59,496 - freqtrade.resolvers.strategy_resolver - INFO - Strategy using trailing_only_offset_is_reached: False
+2025-04-29 01:54:59,496 - freqtrade.resolvers.strategy_resolver - INFO - Strategy using use_custom_stoploss: False
+2025-04-29 01:54:59,497 - freqtrade.resolvers.strategy_resolver - INFO - Strategy using process_only_new_candles: True
+2025-04-29 01:54:59,497 - freqtrade.resolvers.strategy_resolver - INFO - Strategy using order_types: {'entry': 'limit', 'exit': 'limit', 'stoploss': 'limit', 'stoploss_on_exchange': False, 
+'stoploss_on_exchange_interval': 60}
+2025-04-29 01:54:59,497 - freqtrade.resolvers.strategy_resolver - INFO - Strategy using order_time_in_force: {'entry': 'GTC', 'exit': 'GTC'}
+2025-04-29 01:54:59,498 - freqtrade.resolvers.strategy_resolver - INFO - Strategy using stake_currency: USDT
+2025-04-29 01:54:59,498 - freqtrade.resolvers.strategy_resolver - INFO - Strategy using stake_amount: 150
+2025-04-29 01:54:59,498 - freqtrade.resolvers.strategy_resolver - INFO - Strategy using startup_candle_count: 30
+2025-04-29 01:54:59,499 - freqtrade.resolvers.strategy_resolver - INFO - Strategy using unfilledtimeout: {'entry': 5, 'exit': 15, 'exit_timeout_count': 0, 'unit': 'minutes'}
+2025-04-29 01:54:59,499 - freqtrade.resolvers.strategy_resolver - INFO - Strategy using use_exit_signal: True
+2025-04-29 01:54:59,499 - freqtrade.resolvers.strategy_resolver - INFO - Strategy using exit_profit_only: False
+2025-04-29 01:54:59,500 - freqtrade.resolvers.strategy_resolver - INFO - Strategy using ignore_roi_if_entry_signal: False
+2025-04-29 01:54:59,500 - freqtrade.resolvers.strategy_resolver - INFO - Strategy using exit_profit_offset: 0.0
+2025-04-29 01:54:59,500 - freqtrade.resolvers.strategy_resolver - INFO - Strategy using disable_dataframe_checks: False
+2025-04-29 01:54:59,500 - freqtrade.resolvers.strategy_resolver - INFO - Strategy using ignore_buying_expired_candle_after: 0
+2025-04-29 01:54:59,501 - freqtrade.resolvers.strategy_resolver - INFO - Strategy using position_adjustment_enable: False
+2025-04-29 01:54:59,501 - freqtrade.resolvers.strategy_resolver - INFO - Strategy using max_entry_position_adjustment: -1
+2025-04-29 01:54:59,501 - freqtrade.resolvers.strategy_resolver - INFO - Strategy using max_open_trades: 4
+2025-04-29 01:54:59,502 - freqtrade.configuration.config_validation - INFO - Validating configuration ...
+2025-04-29 01:54:59,505 - freqtrade.resolvers.iresolver - INFO - Using resolved pairlist StaticPairList from '/freqtrade/freqtrade/plugins/pairlist/StaticPairList.py'...
+2025-04-29 01:54:59,512 - freqtrade.optimize.backtesting - INFO - Using fee 0.1500% - worst case fee from exchange (lowest tier).
+2025-04-29 01:54:59,512 - freqtrade.data.dataprovider - INFO - Increasing startup_candle_count for freqai on 3m to 14450
+2025-04-29 01:54:59,513 - freqtrade.data.history.history_utils - INFO - Using indicator startup period: 14450 ...
+2025-04-29 01:54:59,672 - freqtrade.optimize.backtesting - INFO - Loading data from 2024-12-01 21:30:00 up to 2025-04-20 00:00:00 (139 days).
+2025-04-29 01:54:59,672 - freqtrade.optimize.backtesting - INFO - Dataload complete. Calculating indicators
+2025-04-29 01:54:59,673 - freqtrade.optimize.backtesting - INFO - Running backtesting for Strategy FreqaiExampleStrategy
+2025-04-29 01:55:01,274 - matplotlib.font_manager - INFO - generated new fontManager
+2025-04-29 01:55:01,489 - freqtrade.resolvers.iresolver - INFO - Using resolved freqaimodel XGBoostRegressor from '/freqtrade/freqtrade/freqai/prediction_models/XGBoostRegressor.py'...
+2025-04-29 01:55:01,490 - freqtrade.freqai.data_drawer - INFO - Could not find existing datadrawer, starting from scratch
+2025-04-29 01:55:01,491 - freqtrade.freqai.data_drawer - INFO - Could not find existing historic_predictions, starting from scratch
+2025-04-29 01:55:01,491 - freqtrade.freqai.freqai_interface - INFO - Set fresh train queue from whitelist. Queue: ['BTC/USDT', 'SOL/USDT']
+2025-04-29 01:55:01,492 - freqtrade.strategy.hyper - INFO - Strategy Parameter: buy_rsi = 39.92672300850069
+2025-04-29 01:55:01,492 - freqtrade.strategy.hyper - INFO - Strategy Parameter: sell_rsi = 69.92672300850067
+2025-04-29 01:55:01,493 - freqtrade.strategy.hyper - INFO - No params for protection found, using default values.
+2025-04-29 01:55:01,498 - FreqaiExampleStrategy - INFO - 处理交易对：BTC/USDT
+2025-04-29 01:55:01,500 - freqtrade.freqai.freqai_interface - INFO - Training 11 timeranges
+2025-04-29 01:55:01,501 - freqtrade.freqai.freqai_interface - INFO - Training BTC/USDT, 1/2 pairs from 2024-12-02 00:00:00 to 2025-01-01 00:00:00, 1/11 trains
+2025-04-29 01:55:01,502 - freqtrade.freqai.data_kitchen - INFO - Could not find backtesting prediction file at 
+/freqtrade/user_data/models/test175/backtesting_predictions/cb_btc_1735689600_prediction.feather
+2025-04-29 01:55:01,602 - freqtrade.data.dataprovider - INFO - Increasing startup_candle_count for freqai on 5m to 8690
+2025-04-29 01:55:01,603 - freqtrade.data.dataprovider - INFO - Loading data for BTC/USDT 5m from 2024-12-01 19:50:00 to 2025-04-20 00:00:00
+2025-04-29 01:55:01,705 - freqtrade.data.dataprovider - INFO - Increasing startup_candle_count for freqai on 1h to 770
+2025-04-29 01:55:01,706 - freqtrade.data.dataprovider - INFO - Loading data for BTC/USDT 1h from 2024-11-29 22:00:00 to 2025-04-20 00:00:00
+2025-04-29 01:55:01,814 - freqtrade.data.dataprovider - INFO - Increasing startup_candle_count for freqai on 3m to 14450
+2025-04-29 01:55:01,815 - freqtrade.data.dataprovider - INFO - Loading data for ETH/USDT 3m from 2024-12-01 21:30:00 to 2025-04-20 00:00:00
+2025-04-29 01:55:01,942 - freqtrade.data.dataprovider - INFO - Increasing startup_candle_count for freqai on 5m to 8690
+2025-04-29 01:55:01,943 - freqtrade.data.dataprovider - INFO - Loading data for ETH/USDT 5m from 2024-12-01 19:50:00 to 2025-04-20 00:00:00
+2025-04-29 01:55:02,037 - freqtrade.data.dataprovider - INFO - Increasing startup_candle_count for freqai on 1h to 770
+2025-04-29 01:55:02,038 - freqtrade.data.dataprovider - INFO - Loading data for ETH/USDT 1h from 2024-11-29 22:00:00 to 2025-04-20 00:00:00
+2025-04-29 01:55:02,113 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：BTC/USDT
+2025-04-29 01:55:02,118 - FreqaiExampleStrategy - INFO - 目标列形状：(14450,)
+2025-04-29 01:55:02,121 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.003572  50.152831
+1    0.003285  50.152831
+2    0.001898  50.152831
+3    0.000484  50.152831
+4    0.001688  50.152831
+2025-04-29 01:55:02,123 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：BTC/USDT
+2025-04-29 01:55:02,129 - FreqaiExampleStrategy - INFO - 目标列形状：(19250,)
+2025-04-29 01:55:02,130 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.003572  50.202701
+1    0.003285  50.202701
+2    0.001898  50.202701
+3    0.000484  50.202701
+4    0.001688  50.202701
+2025-04-29 01:55:02,134 - freqtrade.freqai.freqai_interface - INFO - Could not find model at /freqtrade/user_data/models/test175/sub-train-BTC_1735689600/cb_btc_1735689600
+2025-04-29 01:55:02,135 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Starting training BTC/USDT --------------------
+2025-04-29 01:55:02,151 - freqtrade.freqai.data_kitchen - INFO - BTC/USDT: dropped 0 training points due to NaNs in populated dataset 14400.
+2025-04-29 01:55:02,152 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Training on data from 2024-12-02 to 2024-12-31 --------------------
+2025-04-29 01:55:07,277 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 75 features
+2025-04-29 01:55:07,278 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 11520 data points
+2025-04-29 01:55:08,221 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Done training BTC/USDT (6.09 secs) --------------------
+2025-04-29 01:55:08,222 - freqtrade.freqai.freqai_interface - INFO - Saving metadata to disk.
+2025-04-29 01:55:08,903 - freqtrade.freqai.freqai_interface - INFO - Training BTC/USDT, 1/2 pairs from 2024-12-12 00:00:00 to 2025-01-11 00:00:00, 2/11 trains
+2025-04-29 01:55:08,904 - freqtrade.freqai.data_kitchen - INFO - Could not find backtesting prediction file at 
+/freqtrade/user_data/models/test175/backtesting_predictions/cb_btc_1736553600_prediction.feather
+2025-04-29 01:55:08,907 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：BTC/USDT
+2025-04-29 01:55:08,912 - FreqaiExampleStrategy - INFO - 目标列形状：(19250,)
+2025-04-29 01:55:08,914 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.003572  50.202701
+1    0.003285  50.202701
+2    0.001898  50.202701
+3    0.000484  50.202701
+4    0.001688  50.202701
+2025-04-29 01:55:08,917 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：BTC/USDT
+2025-04-29 01:55:08,924 - FreqaiExampleStrategy - INFO - 目标列形状：(24050,)
+2025-04-29 01:55:08,925 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.003572  50.367593
+1    0.003285  50.367593
+2    0.001898  50.367593
+3    0.000484  50.367593
+4    0.001688  50.367593
+2025-04-29 01:55:08,929 - freqtrade.freqai.freqai_interface - INFO - Could not find model at /freqtrade/user_data/models/test175/sub-train-BTC_1736553600/cb_btc_1736553600
+2025-04-29 01:55:08,930 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Starting training BTC/USDT --------------------
+2025-04-29 01:55:08,946 - freqtrade.freqai.data_kitchen - INFO - BTC/USDT: dropped 0 training points due to NaNs in populated dataset 14400.
+2025-04-29 01:55:08,947 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Training on data from 2024-12-12 to 2025-01-10 --------------------
+2025-04-29 01:55:13,908 - datasieve.pipeline - INFO - DI tossed 5 predictions for being too far from training data.
+2025-04-29 01:55:13,911 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 75 features
+2025-04-29 01:55:13,912 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 11520 data points
+2025-04-29 01:55:14,692 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Done training BTC/USDT (5.76 secs) --------------------
+2025-04-29 01:55:14,693 - freqtrade.freqai.freqai_interface - INFO - Saving metadata to disk.
+2025-04-29 01:55:15,250 - freqtrade.freqai.freqai_interface - INFO - Training BTC/USDT, 1/2 pairs from 2024-12-22 00:00:00 to 2025-01-21 00:00:00, 3/11 trains
+2025-04-29 01:55:15,250 - freqtrade.freqai.data_kitchen - INFO - Could not find backtesting prediction file at 
+/freqtrade/user_data/models/test175/backtesting_predictions/cb_btc_1737417600_prediction.feather
+2025-04-29 01:55:15,254 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：BTC/USDT
+2025-04-29 01:55:15,261 - FreqaiExampleStrategy - INFO - 目标列形状：(24050,)
+2025-04-29 01:55:15,262 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.003572  50.367593
+1    0.003285  50.367593
+2    0.001898  50.367593
+3    0.000484  50.367593
+4    0.001688  50.367593
+2025-04-29 01:55:15,268 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：BTC/USDT
+2025-04-29 01:55:15,275 - FreqaiExampleStrategy - INFO - 目标列形状：(28850,)
+2025-04-29 01:55:15,276 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.003572  50.305589
+1    0.003285  50.305589
+2    0.001898  50.305589
+3    0.000484  50.305589
+4    0.001688  50.305589
+2025-04-29 01:55:15,281 - freqtrade.freqai.freqai_interface - INFO - Could not find model at /freqtrade/user_data/models/test175/sub-train-BTC_1737417600/cb_btc_1737417600
+2025-04-29 01:55:15,281 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Starting training BTC/USDT --------------------
+2025-04-29 01:55:15,297 - freqtrade.freqai.data_kitchen - INFO - BTC/USDT: dropped 0 training points due to NaNs in populated dataset 14400.
+2025-04-29 01:55:15,298 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Training on data from 2024-12-22 to 2025-01-20 --------------------
+2025-04-29 01:55:20,324 - datasieve.pipeline - INFO - DI tossed 1622 predictions for being too far from training data.
+2025-04-29 01:55:20,327 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 75 features
+2025-04-29 01:55:20,327 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 11520 data points
+2025-04-29 01:55:21,007 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Done training BTC/USDT (5.73 secs) --------------------
+2025-04-29 01:55:21,008 - freqtrade.freqai.freqai_interface - INFO - Saving metadata to disk.
+2025-04-29 01:55:21,504 - freqtrade.freqai.freqai_interface - INFO - Training BTC/USDT, 1/2 pairs from 2025-01-01 00:00:00 to 2025-01-31 00:00:00, 4/11 trains
+2025-04-29 01:55:21,505 - freqtrade.freqai.data_kitchen - INFO - Could not find backtesting prediction file at 
+/freqtrade/user_data/models/test175/backtesting_predictions/cb_btc_1738281600_prediction.feather
+2025-04-29 01:55:21,510 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：BTC/USDT
+2025-04-29 01:55:21,516 - FreqaiExampleStrategy - INFO - 目标列形状：(28850,)
+2025-04-29 01:55:21,517 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.003572  50.305589
+1    0.003285  50.305589
+2    0.001898  50.305589
+3    0.000484  50.305589
+4    0.001688  50.305589
+2025-04-29 01:55:21,522 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：BTC/USDT
+2025-04-29 01:55:21,528 - FreqaiExampleStrategy - INFO - 目标列形状：(33650,)
+2025-04-29 01:55:21,529 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.003572  50.168798
+1    0.003285  50.168798
+2    0.001898  50.168798
+3    0.000484  50.168798
+4    0.001688  50.168798
+2025-04-29 01:55:21,533 - freqtrade.freqai.freqai_interface - INFO - Could not find model at /freqtrade/user_data/models/test175/sub-train-BTC_1738281600/cb_btc_1738281600
+2025-04-29 01:55:21,534 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Starting training BTC/USDT --------------------
+2025-04-29 01:55:21,550 - freqtrade.freqai.data_kitchen - INFO - BTC/USDT: dropped 0 training points due to NaNs in populated dataset 14400.
+2025-04-29 01:55:21,550 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Training on data from 2025-01-01 to 2025-01-30 --------------------
+2025-04-29 01:55:26,605 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 75 features
+2025-04-29 01:55:26,606 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 11520 data points
+2025-04-29 01:55:27,556 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Done training BTC/USDT (6.02 secs) --------------------
+2025-04-29 01:55:27,557 - freqtrade.freqai.freqai_interface - INFO - Saving metadata to disk.
+2025-04-29 01:55:28,076 - freqtrade.freqai.freqai_interface - INFO - Training BTC/USDT, 1/2 pairs from 2025-01-11 00:00:00 to 2025-02-10 00:00:00, 5/11 trains
+2025-04-29 01:55:28,077 - freqtrade.freqai.data_kitchen - INFO - Could not find backtesting prediction file at 
+/freqtrade/user_data/models/test175/backtesting_predictions/cb_btc_1739145600_prediction.feather
+2025-04-29 01:55:28,081 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：BTC/USDT
+2025-04-29 01:55:28,088 - FreqaiExampleStrategy - INFO - 目标列形状：(33650,)
+2025-04-29 01:55:28,089 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.003572  50.168798
+1    0.003285  50.168798
+2    0.001898  50.168798
+3    0.000484  50.168798
+4    0.001688  50.168798
+2025-04-29 01:55:28,094 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：BTC/USDT
+2025-04-29 01:55:28,100 - FreqaiExampleStrategy - INFO - 目标列形状：(38450,)
+2025-04-29 01:55:28,102 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.003572  50.167897
+1    0.003285  50.167897
+2    0.001898  50.167897
+3    0.000484  50.167897
+4    0.001688  50.167897
+2025-04-29 01:55:28,106 - freqtrade.freqai.freqai_interface - INFO - Could not find model at /freqtrade/user_data/models/test175/sub-train-BTC_1739145600/cb_btc_1739145600
+2025-04-29 01:55:28,107 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Starting training BTC/USDT --------------------
+2025-04-29 01:55:28,123 - freqtrade.freqai.data_kitchen - INFO - BTC/USDT: dropped 0 training points due to NaNs in populated dataset 14400.
+2025-04-29 01:55:28,124 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Training on data from 2025-01-11 to 2025-02-09 --------------------
+2025-04-29 01:55:33,123 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 75 features
+2025-04-29 01:55:33,124 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 11520 data points
+2025-04-29 01:55:33,929 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Done training BTC/USDT (5.82 secs) --------------------
+2025-04-29 01:55:33,930 - freqtrade.freqai.freqai_interface - INFO - Saving metadata to disk.
+2025-04-29 01:55:34,433 - freqtrade.freqai.freqai_interface - INFO - Training BTC/USDT, 1/2 pairs from 2025-01-21 00:00:00 to 2025-02-20 00:00:00, 6/11 trains
+2025-04-29 01:55:34,434 - freqtrade.freqai.data_kitchen - INFO - Could not find backtesting prediction file at 
+/freqtrade/user_data/models/test175/backtesting_predictions/cb_btc_1740009600_prediction.feather
+2025-04-29 01:55:34,440 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：BTC/USDT
+2025-04-29 01:55:34,447 - FreqaiExampleStrategy - INFO - 目标列形状：(38450,)
+2025-04-29 01:55:34,448 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.003572  50.167897
+1    0.003285  50.167897
+2    0.001898  50.167897
+3    0.000484  50.167897
+4    0.001688  50.167897
+2025-04-29 01:55:34,453 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：BTC/USDT
+2025-04-29 01:55:34,459 - FreqaiExampleStrategy - INFO - 目标列形状：(43250,)
+2025-04-29 01:55:34,461 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.003572  50.107698
+1    0.003285  50.107698
+2    0.001898  50.107698
+3    0.000484  50.107698
+4    0.001688  50.107698
+2025-04-29 01:55:34,465 - freqtrade.freqai.freqai_interface - INFO - Could not find model at /freqtrade/user_data/models/test175/sub-train-BTC_1740009600/cb_btc_1740009600
+2025-04-29 01:55:34,466 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Starting training BTC/USDT --------------------
+2025-04-29 01:55:34,482 - freqtrade.freqai.data_kitchen - INFO - BTC/USDT: dropped 0 training points due to NaNs in populated dataset 14400.
+2025-04-29 01:55:34,483 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Training on data from 2025-01-21 to 2025-02-19 --------------------
+2025-04-29 01:55:39,369 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 75 features
+2025-04-29 01:55:39,370 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 11520 data points
+2025-04-29 01:55:40,266 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Done training BTC/USDT (5.80 secs) --------------------
+2025-04-29 01:55:40,267 - freqtrade.freqai.freqai_interface - INFO - Saving metadata to disk.
+2025-04-29 01:55:40,801 - freqtrade.freqai.freqai_interface - INFO - Training BTC/USDT, 1/2 pairs from 2025-01-31 00:00:00 to 2025-03-02 00:00:00, 7/11 trains
+2025-04-29 01:55:40,802 - freqtrade.freqai.data_kitchen - INFO - Could not find backtesting prediction file at 
+/freqtrade/user_data/models/test175/backtesting_predictions/cb_btc_1740873600_prediction.feather
+2025-04-29 01:55:40,807 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：BTC/USDT
+2025-04-29 01:55:40,814 - FreqaiExampleStrategy - INFO - 目标列形状：(43250,)
+2025-04-29 01:55:40,816 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.003572  50.107698
+1    0.003285  50.107698
+2    0.001898  50.107698
+3    0.000484  50.107698
+4    0.001688  50.107698
+2025-04-29 01:55:40,821 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：BTC/USDT
+2025-04-29 01:55:40,827 - FreqaiExampleStrategy - INFO - 目标列形状：(48050,)
+2025-04-29 01:55:40,829 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.003572  50.079166
+1    0.003285  50.079166
+2    0.001898  50.079166
+3    0.000484  50.079166
+4    0.001688  50.079166
+2025-04-29 01:55:40,833 - freqtrade.freqai.freqai_interface - INFO - Could not find model at /freqtrade/user_data/models/test175/sub-train-BTC_1740873600/cb_btc_1740873600
+2025-04-29 01:55:40,834 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Starting training BTC/USDT --------------------
+2025-04-29 01:55:40,849 - freqtrade.freqai.data_kitchen - INFO - BTC/USDT: dropped 0 training points due to NaNs in populated dataset 14400.
+2025-04-29 01:55:40,850 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Training on data from 2025-01-31 to 2025-03-01 --------------------
+2025-04-29 01:55:45,643 - datasieve.pipeline - INFO - DI tossed 2275 predictions for being too far from training data.
+2025-04-29 01:55:45,646 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 75 features
+2025-04-29 01:55:45,647 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 11520 data points
+2025-04-29 01:55:46,544 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Done training BTC/USDT (5.71 secs) --------------------
+2025-04-29 01:55:46,544 - freqtrade.freqai.freqai_interface - INFO - Saving metadata to disk.
+2025-04-29 01:55:47,092 - freqtrade.freqai.freqai_interface - INFO - Training BTC/USDT, 1/2 pairs from 2025-02-10 00:00:00 to 2025-03-12 00:00:00, 8/11 trains
+2025-04-29 01:55:47,092 - freqtrade.freqai.data_kitchen - INFO - Could not find backtesting prediction file at 
+/freqtrade/user_data/models/test175/backtesting_predictions/cb_btc_1741737600_prediction.feather
+2025-04-29 01:55:47,100 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：BTC/USDT
+2025-04-29 01:55:47,107 - FreqaiExampleStrategy - INFO - 目标列形状：(48050,)
+2025-04-29 01:55:47,109 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.003572  50.079166
+1    0.003285  50.079166
+2    0.001898  50.079166
+3    0.000484  50.079166
+4    0.001688  50.079166
+2025-04-29 01:55:47,115 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：BTC/USDT
+2025-04-29 01:55:47,122 - FreqaiExampleStrategy - INFO - 目标列形状：(52850,)
+2025-04-29 01:55:47,123 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.003572  50.102027
+1    0.003285  50.102027
+2    0.001898  50.102027
+3    0.000484  50.102027
+4    0.001688  50.102027
+2025-04-29 01:55:47,128 - freqtrade.freqai.freqai_interface - INFO - Could not find model at /freqtrade/user_data/models/test175/sub-train-BTC_1741737600/cb_btc_1741737600
+2025-04-29 01:55:47,129 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Starting training BTC/USDT --------------------
+2025-04-29 01:55:47,145 - freqtrade.freqai.data_kitchen - INFO - BTC/USDT: dropped 0 training points due to NaNs in populated dataset 14400.
+2025-04-29 01:55:47,145 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Training on data from 2025-02-10 to 2025-03-11 --------------------
+2025-04-29 01:55:51,987 - datasieve.pipeline - INFO - DI tossed 18 predictions for being too far from training data.
+2025-04-29 01:55:51,989 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 75 features
+2025-04-29 01:55:51,989 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 11520 data points
+2025-04-29 01:55:52,741 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Done training BTC/USDT (5.61 secs) --------------------
+2025-04-29 01:55:52,742 - freqtrade.freqai.freqai_interface - INFO - Saving metadata to disk.
+2025-04-29 01:55:53,285 - freqtrade.freqai.freqai_interface - INFO - Training BTC/USDT, 1/2 pairs from 2025-02-20 00:00:00 to 2025-03-22 00:00:00, 9/11 trains
+2025-04-29 01:55:53,286 - freqtrade.freqai.data_kitchen - INFO - Could not find backtesting prediction file at 
+/freqtrade/user_data/models/test175/backtesting_predictions/cb_btc_1742601600_prediction.feather
+2025-04-29 01:55:53,291 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：BTC/USDT
+2025-04-29 01:55:53,298 - FreqaiExampleStrategy - INFO - 目标列形状：(52850,)
+2025-04-29 01:55:53,300 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.003572  50.102027
+1    0.003285  50.102027
+2    0.001898  50.102027
+3    0.000484  50.102027
+4    0.001688  50.102027
+2025-04-29 01:55:53,309 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：BTC/USDT
+2025-04-29 01:55:53,316 - FreqaiExampleStrategy - INFO - 目标列形状：(57650,)
+2025-04-29 01:55:53,318 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.003572  50.079967
+1    0.003285  50.079967
+2    0.001898  50.079967
+3    0.000484  50.079967
+4    0.001688  50.079967
+2025-04-29 01:55:53,322 - freqtrade.freqai.freqai_interface - INFO - Could not find model at /freqtrade/user_data/models/test175/sub-train-BTC_1742601600/cb_btc_1742601600
+2025-04-29 01:55:53,323 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Starting training BTC/USDT --------------------
+2025-04-29 01:55:53,339 - freqtrade.freqai.data_kitchen - INFO - BTC/USDT: dropped 0 training points due to NaNs in populated dataset 14400.
+2025-04-29 01:55:53,340 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Training on data from 2025-02-20 to 2025-03-21 --------------------
+2025-04-29 01:55:58,184 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 75 features
+2025-04-29 01:55:58,185 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 11520 data points
+2025-04-29 01:55:59,097 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Done training BTC/USDT (5.77 secs) --------------------
+2025-04-29 01:55:59,098 - freqtrade.freqai.freqai_interface - INFO - Saving metadata to disk.
+2025-04-29 01:55:59,706 - freqtrade.freqai.freqai_interface - INFO - Training BTC/USDT, 1/2 pairs from 2025-03-02 00:00:00 to 2025-04-01 00:00:00, 10/11 trains
+2025-04-29 01:55:59,706 - freqtrade.freqai.data_kitchen - INFO - Could not find backtesting prediction file at 
+/freqtrade/user_data/models/test175/backtesting_predictions/cb_btc_1743465600_prediction.feather
+2025-04-29 01:55:59,715 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：BTC/USDT
+2025-04-29 01:55:59,723 - FreqaiExampleStrategy - INFO - 目标列形状：(57650,)
+2025-04-29 01:55:59,725 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.003572  50.079967
+1    0.003285  50.079967
+2    0.001898  50.079967
+3    0.000484  50.079967
+4    0.001688  50.079967
+2025-04-29 01:55:59,732 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：BTC/USDT
+2025-04-29 01:55:59,739 - FreqaiExampleStrategy - INFO - 目标列形状：(62450,)
+2025-04-29 01:55:59,741 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.003572  50.024153
+1    0.003285  50.024153
+2    0.001898  50.024153
+3    0.000484  50.024153
+4    0.001688  50.024153
+2025-04-29 01:55:59,745 - freqtrade.freqai.freqai_interface - INFO - Could not find model at /freqtrade/user_data/models/test175/sub-train-BTC_1743465600/cb_btc_1743465600
+2025-04-29 01:55:59,746 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Starting training BTC/USDT --------------------
+2025-04-29 01:55:59,762 - freqtrade.freqai.data_kitchen - INFO - BTC/USDT: dropped 0 training points due to NaNs in populated dataset 14400.
+2025-04-29 01:55:59,762 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Training on data from 2025-03-02 to 2025-03-31 --------------------
+2025-04-29 01:56:04,571 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 75 features
+2025-04-29 01:56:04,571 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 11520 data points
+2025-04-29 01:56:05,520 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Done training BTC/USDT (5.77 secs) --------------------
+2025-04-29 01:56:05,521 - freqtrade.freqai.freqai_interface - INFO - Saving metadata to disk.
+2025-04-29 01:56:06,027 - freqtrade.freqai.freqai_interface - INFO - Training BTC/USDT, 1/2 pairs from 2025-03-12 00:00:00 to 2025-04-11 00:00:00, 11/11 trains
+2025-04-29 01:56:06,027 - freqtrade.freqai.data_kitchen - INFO - Could not find backtesting prediction file at 
+/freqtrade/user_data/models/test175/backtesting_predictions/cb_btc_1744329600_prediction.feather
+2025-04-29 01:56:06,037 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：BTC/USDT
+2025-04-29 01:56:06,045 - FreqaiExampleStrategy - INFO - 目标列形状：(62450,)
+2025-04-29 01:56:06,046 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.003572  50.024153
+1    0.003285  50.024153
+2    0.001898  50.024153
+3    0.000484  50.024153
+4    0.001688  50.024153
+2025-04-29 01:56:06,057 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：BTC/USDT
+2025-04-29 01:56:06,064 - FreqaiExampleStrategy - INFO - 目标列形状：(66770,)
+2025-04-29 01:56:06,065 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.003572  50.093162
+1    0.003285  50.093162
+2    0.001898  50.093162
+3    0.000484  50.093162
+4    0.001688  50.093162
+2025-04-29 01:56:06,070 - freqtrade.freqai.freqai_interface - INFO - Could not find model at /freqtrade/user_data/models/test175/sub-train-BTC_1744329600/cb_btc_1744329600
+2025-04-29 01:56:06,071 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Starting training BTC/USDT --------------------
+2025-04-29 01:56:06,087 - freqtrade.freqai.data_kitchen - INFO - BTC/USDT: dropped 0 training points due to NaNs in populated dataset 14400.
+2025-04-29 01:56:06,088 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Training on data from 2025-03-12 to 2025-04-10 --------------------
+2025-04-29 01:56:10,904 - datasieve.pipeline - INFO - DI tossed 2001 predictions for being too far from training data.
+2025-04-29 01:56:10,907 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 75 features
+2025-04-29 01:56:10,907 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 11520 data points
+2025-04-29 01:56:11,705 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Done training BTC/USDT (5.63 secs) --------------------
+2025-04-29 01:56:11,706 - freqtrade.freqai.freqai_interface - INFO - Saving metadata to disk.
+2025-04-29 01:56:12,255 - FreqaiExampleStrategy - INFO - 动态参数：buy_rsi=39.26145316407591, sell_rsi=59.26145316407591, stoploss=-0.15, trailing_stop_positive=0.05
+2025-04-29 01:56:12,275 - FreqaiExampleStrategy - INFO - up_or_down 值统计：
+up_or_down
+1    33535
+0    33236
+2025-04-29 01:56:12,276 - FreqaiExampleStrategy - INFO - do_predict 值统计：
+do_predict
+0.0    35773
+1.0    30998
+2025-04-29 01:56:12,279 - FreqaiExampleStrategy - INFO - 处理交易对：SOL/USDT
+2025-04-29 01:56:12,281 - freqtrade.freqai.freqai_interface - INFO - Training 11 timeranges
+2025-04-29 01:56:12,282 - freqtrade.freqai.freqai_interface - INFO - Training SOL/USDT, 2/2 pairs from 2024-12-02 00:00:00 to 2025-01-01 00:00:00, 1/11 trains
+2025-04-29 01:56:12,283 - freqtrade.freqai.data_kitchen - INFO - Could not find backtesting prediction file at 
+/freqtrade/user_data/models/test175/backtesting_predictions/cb_sol_1735689600_prediction.feather
+2025-04-29 01:56:12,334 - freqtrade.data.dataprovider - INFO - Increasing startup_candle_count for freqai on 5m to 8690
+2025-04-29 01:56:12,335 - freqtrade.data.dataprovider - INFO - Loading data for SOL/USDT 5m from 2024-12-01 19:50:00 to 2025-04-20 00:00:00
+2025-04-29 01:56:12,422 - freqtrade.data.dataprovider - INFO - Increasing startup_candle_count for freqai on 1h to 770
+2025-04-29 01:56:12,422 - freqtrade.data.dataprovider - INFO - Loading data for SOL/USDT 1h from 2024-11-29 22:00:00 to 2025-04-20 00:00:00
+2025-04-29 01:56:12,518 - freqtrade.data.dataprovider - INFO - Increasing startup_candle_count for freqai on 3m to 14450
+2025-04-29 01:56:12,519 - freqtrade.data.dataprovider - INFO - Loading data for BTC/USDT 3m from 2024-12-01 21:30:00 to 2025-04-20 00:00:00
+2025-04-29 01:56:13,040 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：SOL/USDT
+2025-04-29 01:56:13,046 - FreqaiExampleStrategy - INFO - 目标列形状：(14450,)
+2025-04-29 01:56:13,047 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.002704   49.58814
+1    0.003044   49.58814
+2    0.000465   49.58814
+3   -0.000380   49.58814
+4    0.002829   49.58814
+2025-04-29 01:56:13,052 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：SOL/USDT
+2025-04-29 01:56:13,057 - FreqaiExampleStrategy - INFO - 目标列形状：(19250,)
+2025-04-29 01:56:13,059 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.002704   49.68088
+1    0.003044   49.68088
+2    0.000465   49.68088
+3   -0.000380   49.68088
+4    0.002829   49.68088
+2025-04-29 01:56:13,066 - freqtrade.freqai.freqai_interface - INFO - Could not find model at /freqtrade/user_data/models/test175/sub-train-SOL_1735689600/cb_sol_1735689600
+2025-04-29 01:56:13,066 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Starting training SOL/USDT --------------------
+2025-04-29 01:56:13,095 - freqtrade.freqai.data_kitchen - INFO - SOL/USDT: dropped 0 training points due to NaNs in populated dataset 14400.
+2025-04-29 01:56:13,096 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Training on data from 2024-12-02 to 2024-12-31 --------------------
+2025-04-29 01:56:18,126 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 111 features
+2025-04-29 01:56:18,126 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 11520 data points
+2025-04-29 01:56:19,586 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Done training SOL/USDT (6.52 secs) --------------------
+2025-04-29 01:56:19,587 - freqtrade.freqai.freqai_interface - INFO - Saving metadata to disk.
+2025-04-29 01:56:20,174 - freqtrade.freqai.freqai_interface - INFO - Training SOL/USDT, 2/2 pairs from 2024-12-12 00:00:00 to 2025-01-11 00:00:00, 2/11 trains
+2025-04-29 01:56:20,175 - freqtrade.freqai.data_kitchen - INFO - Could not find backtesting prediction file at 
+/freqtrade/user_data/models/test175/backtesting_predictions/cb_sol_1736553600_prediction.feather
+2025-04-29 01:56:20,179 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：SOL/USDT
+2025-04-29 01:56:20,185 - FreqaiExampleStrategy - INFO - 目标列形状：(19250,)
+2025-04-29 01:56:20,186 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.002704   49.68088
+1    0.003044   49.68088
+2    0.000465   49.68088
+3   -0.000380   49.68088
+4    0.002829   49.68088
+2025-04-29 01:56:20,192 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：SOL/USDT
+2025-04-29 01:56:20,197 - FreqaiExampleStrategy - INFO - 目标列形状：(24050,)
+2025-04-29 01:56:20,199 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.002704   49.97721
+1    0.003044   49.97721
+2    0.000465   49.97721
+3   -0.000380   49.97721
+4    0.002829   49.97721
+2025-04-29 01:56:20,204 - freqtrade.freqai.freqai_interface - INFO - Could not find model at /freqtrade/user_data/models/test175/sub-train-SOL_1736553600/cb_sol_1736553600
+2025-04-29 01:56:20,205 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Starting training SOL/USDT --------------------
+2025-04-29 01:56:20,227 - freqtrade.freqai.data_kitchen - INFO - SOL/USDT: dropped 0 training points due to NaNs in populated dataset 14400.
+2025-04-29 01:56:20,228 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Training on data from 2024-12-12 to 2025-01-10 --------------------
+2025-04-29 01:56:25,109 - datasieve.pipeline - INFO - DI tossed 5 predictions for being too far from training data.
+2025-04-29 01:56:25,112 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 111 features
+2025-04-29 01:56:25,112 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 11520 data points
+2025-04-29 01:56:26,510 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Done training SOL/USDT (6.30 secs) --------------------
+2025-04-29 01:56:26,511 - freqtrade.freqai.freqai_interface - INFO - Saving metadata to disk.
+2025-04-29 01:56:27,072 - freqtrade.freqai.freqai_interface - INFO - Training SOL/USDT, 2/2 pairs from 2024-12-22 00:00:00 to 2025-01-21 00:00:00, 3/11 trains
+2025-04-29 01:56:27,073 - freqtrade.freqai.data_kitchen - INFO - Could not find backtesting prediction file at 
+/freqtrade/user_data/models/test175/backtesting_predictions/cb_sol_1737417600_prediction.feather
+2025-04-29 01:56:27,079 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：SOL/USDT
+2025-04-29 01:56:27,085 - FreqaiExampleStrategy - INFO - 目标列形状：(24050,)
+2025-04-29 01:56:27,086 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.002704   49.97721
+1    0.003044   49.97721
+2    0.000465   49.97721
+3   -0.000380   49.97721
+4    0.002829   49.97721
+2025-04-29 01:56:27,094 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：SOL/USDT
+2025-04-29 01:56:27,100 - FreqaiExampleStrategy - INFO - 目标列形状：(28850,)
+2025-04-29 01:56:27,102 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.002704  49.941408
+1    0.003044  49.941408
+2    0.000465  49.941408
+3   -0.000380  49.941408
+4    0.002829  49.941408
+2025-04-29 01:56:27,108 - freqtrade.freqai.freqai_interface - INFO - Could not find model at /freqtrade/user_data/models/test175/sub-train-SOL_1737417600/cb_sol_1737417600
+2025-04-29 01:56:27,109 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Starting training SOL/USDT --------------------
+2025-04-29 01:56:27,130 - freqtrade.freqai.data_kitchen - INFO - SOL/USDT: dropped 0 training points due to NaNs in populated dataset 14400.
+2025-04-29 01:56:27,131 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Training on data from 2024-12-22 to 2025-01-20 --------------------
+2025-04-29 01:56:32,206 - datasieve.pipeline - INFO - DI tossed 1523 predictions for being too far from training data.
+2025-04-29 01:56:32,209 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 111 features
+2025-04-29 01:56:32,210 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 11520 data points
+2025-04-29 01:56:33,558 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Done training SOL/USDT (6.45 secs) --------------------
+2025-04-29 01:56:33,558 - freqtrade.freqai.freqai_interface - INFO - Saving metadata to disk.
+2025-04-29 01:56:34,118 - freqtrade.freqai.freqai_interface - INFO - Training SOL/USDT, 2/2 pairs from 2025-01-01 00:00:00 to 2025-01-31 00:00:00, 4/11 trains
+2025-04-29 01:56:34,119 - freqtrade.freqai.data_kitchen - INFO - Could not find backtesting prediction file at 
+/freqtrade/user_data/models/test175/backtesting_predictions/cb_sol_1738281600_prediction.feather
+2025-04-29 01:56:34,124 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：SOL/USDT
+2025-04-29 01:56:34,130 - FreqaiExampleStrategy - INFO - 目标列形状：(28850,)
+2025-04-29 01:56:34,131 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.002704  49.941408
+1    0.003044  49.941408
+2    0.000465  49.941408
+3   -0.000380  49.941408
+4    0.002829  49.941408
+2025-04-29 01:56:34,137 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：SOL/USDT
+2025-04-29 01:56:34,143 - FreqaiExampleStrategy - INFO - 目标列形状：(33650,)
+2025-04-29 01:56:34,144 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.002704  49.830756
+1    0.003044  49.830756
+2    0.000465  49.830756
+3   -0.000380  49.830756
+4    0.002829  49.830756
+2025-04-29 01:56:34,149 - freqtrade.freqai.freqai_interface - INFO - Could not find model at /freqtrade/user_data/models/test175/sub-train-SOL_1738281600/cb_sol_1738281600
+2025-04-29 01:56:34,150 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Starting training SOL/USDT --------------------
+2025-04-29 01:56:34,173 - freqtrade.freqai.data_kitchen - INFO - SOL/USDT: dropped 0 training points due to NaNs in populated dataset 14400.
+2025-04-29 01:56:34,173 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Training on data from 2025-01-01 to 2025-01-30 --------------------
+2025-04-29 01:56:39,271 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 111 features
+2025-04-29 01:56:39,271 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 11520 data points
+2025-04-29 01:56:40,600 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Done training SOL/USDT (6.45 secs) --------------------
+2025-04-29 01:56:40,601 - freqtrade.freqai.freqai_interface - INFO - Saving metadata to disk.
+2025-04-29 01:56:41,171 - freqtrade.freqai.freqai_interface - INFO - Training SOL/USDT, 2/2 pairs from 2025-01-11 00:00:00 to 2025-02-10 00:00:00, 5/11 trains
+2025-04-29 01:56:41,172 - freqtrade.freqai.data_kitchen - INFO - Could not find backtesting prediction file at 
+/freqtrade/user_data/models/test175/backtesting_predictions/cb_sol_1739145600_prediction.feather
+2025-04-29 01:56:41,177 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：SOL/USDT
+2025-04-29 01:56:41,183 - FreqaiExampleStrategy - INFO - 目标列形状：(33650,)
+2025-04-29 01:56:41,185 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.002704  49.830756
+1    0.003044  49.830756
+2    0.000465  49.830756
+3   -0.000380  49.830756
+4    0.002829  49.830756
+2025-04-29 01:56:41,193 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：SOL/USDT
+2025-04-29 01:56:41,200 - FreqaiExampleStrategy - INFO - 目标列形状：(38450,)
+2025-04-29 01:56:41,201 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.002704  49.714422
+1    0.003044  49.714422
+2    0.000465  49.714422
+3   -0.000380  49.714422
+4    0.002829  49.714422
+2025-04-29 01:56:41,206 - freqtrade.freqai.freqai_interface - INFO - Could not find model at /freqtrade/user_data/models/test175/sub-train-SOL_1739145600/cb_sol_1739145600
+2025-04-29 01:56:41,207 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Starting training SOL/USDT --------------------
+2025-04-29 01:56:41,228 - freqtrade.freqai.data_kitchen - INFO - SOL/USDT: dropped 0 training points due to NaNs in populated dataset 14400.
+2025-04-29 01:56:41,229 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Training on data from 2025-01-11 to 2025-02-09 --------------------
+2025-04-29 01:56:46,277 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 111 features
+2025-04-29 01:56:46,278 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 11520 data points
+2025-04-29 01:56:47,778 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Done training SOL/USDT (6.57 secs) --------------------
+2025-04-29 01:56:47,779 - freqtrade.freqai.freqai_interface - INFO - Saving metadata to disk.
+2025-04-29 01:56:48,320 - freqtrade.freqai.freqai_interface - INFO - Training SOL/USDT, 2/2 pairs from 2025-01-21 00:00:00 to 2025-02-20 00:00:00, 6/11 trains
+2025-04-29 01:56:48,321 - freqtrade.freqai.data_kitchen - INFO - Could not find backtesting prediction file at 
+/freqtrade/user_data/models/test175/backtesting_predictions/cb_sol_1740009600_prediction.feather
+2025-04-29 01:56:48,327 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：SOL/USDT
+2025-04-29 01:56:48,333 - FreqaiExampleStrategy - INFO - 目标列形状：(38450,)
+2025-04-29 01:56:48,334 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.002704  49.714422
+1    0.003044  49.714422
+2    0.000465  49.714422
+3   -0.000380  49.714422
+4    0.002829  49.714422
+2025-04-29 01:56:48,346 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：SOL/USDT
+2025-04-29 01:56:48,353 - FreqaiExampleStrategy - INFO - 目标列形状：(43250,)
+2025-04-29 01:56:48,354 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.002704  49.626186
+1    0.003044  49.626186
+2    0.000465  49.626186
+3   -0.000380  49.626186
+4    0.002829  49.626186
+2025-04-29 01:56:48,361 - freqtrade.freqai.freqai_interface - INFO - Could not find model at /freqtrade/user_data/models/test175/sub-train-SOL_1740009600/cb_sol_1740009600
+2025-04-29 01:56:48,361 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Starting training SOL/USDT --------------------
+2025-04-29 01:56:48,383 - freqtrade.freqai.data_kitchen - INFO - SOL/USDT: dropped 0 training points due to NaNs in populated dataset 14400.
+2025-04-29 01:56:48,383 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Training on data from 2025-01-21 to 2025-02-19 --------------------
+2025-04-29 01:56:53,532 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 111 features
+2025-04-29 01:56:53,533 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 11520 data points
+2025-04-29 01:56:54,862 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Done training SOL/USDT (6.50 secs) --------------------
+2025-04-29 01:56:54,863 - freqtrade.freqai.freqai_interface - INFO - Saving metadata to disk.
+2025-04-29 01:56:55,419 - freqtrade.freqai.freqai_interface - INFO - Training SOL/USDT, 2/2 pairs from 2025-01-31 00:00:00 to 2025-03-02 00:00:00, 7/11 trains
+2025-04-29 01:56:55,420 - freqtrade.freqai.data_kitchen - INFO - Could not find backtesting prediction file at 
+/freqtrade/user_data/models/test175/backtesting_predictions/cb_sol_1740873600_prediction.feather
+2025-04-29 01:56:55,426 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：SOL/USDT
+2025-04-29 01:56:55,433 - FreqaiExampleStrategy - INFO - 目标列形状：(43250,)
+2025-04-29 01:56:55,435 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.002704  49.626186
+1    0.003044  49.626186
+2    0.000465  49.626186
+3   -0.000380  49.626186
+4    0.002829  49.626186
+2025-04-29 01:56:55,445 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：SOL/USDT
+2025-04-29 01:56:55,452 - FreqaiExampleStrategy - INFO - 目标列形状：(48050,)
+2025-04-29 01:56:55,453 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.002704  49.568812
+1    0.003044  49.568812
+2    0.000465  49.568812
+3   -0.000380  49.568812
+4    0.002829  49.568812
+2025-04-29 01:56:55,459 - freqtrade.freqai.freqai_interface - INFO - Could not find model at /freqtrade/user_data/models/test175/sub-train-SOL_1740873600/cb_sol_1740873600
+2025-04-29 01:56:55,459 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Starting training SOL/USDT --------------------
+2025-04-29 01:56:55,481 - freqtrade.freqai.data_kitchen - INFO - SOL/USDT: dropped 0 training points due to NaNs in populated dataset 14400.
+2025-04-29 01:56:55,482 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Training on data from 2025-01-31 to 2025-03-01 --------------------
+2025-04-29 01:57:00,566 - datasieve.pipeline - INFO - DI tossed 2417 predictions for being too far from training data.
+2025-04-29 01:57:00,569 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 111 features
+2025-04-29 01:57:00,570 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 11520 data points
+2025-04-29 01:57:02,441 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Done training SOL/USDT (6.98 secs) --------------------
+2025-04-29 01:57:02,442 - freqtrade.freqai.freqai_interface - INFO - Saving metadata to disk.
+2025-04-29 01:57:02,968 - freqtrade.freqai.freqai_interface - INFO - Training SOL/USDT, 2/2 pairs from 2025-02-10 00:00:00 to 2025-03-12 00:00:00, 8/11 trains
+2025-04-29 01:57:02,968 - freqtrade.freqai.data_kitchen - INFO - Could not find backtesting prediction file at 
+/freqtrade/user_data/models/test175/backtesting_predictions/cb_sol_1741737600_prediction.feather
+2025-04-29 01:57:02,980 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：SOL/USDT
+2025-04-29 01:57:02,987 - FreqaiExampleStrategy - INFO - 目标列形状：(48050,)
+2025-04-29 01:57:02,989 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.002704  49.568812
+1    0.003044  49.568812
+2    0.000465  49.568812
+3   -0.000380  49.568812
+4    0.002829  49.568812
+2025-04-29 01:57:03,001 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：SOL/USDT
+2025-04-29 01:57:03,007 - FreqaiExampleStrategy - INFO - 目标列形状：(52850,)
+2025-04-29 01:57:03,009 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.002704  49.623338
+1    0.003044  49.623338
+2    0.000465  49.623338
+3   -0.000380  49.623338
+4    0.002829  49.623338
+2025-04-29 01:57:03,014 - freqtrade.freqai.freqai_interface - INFO - Could not find model at /freqtrade/user_data/models/test175/sub-train-SOL_1741737600/cb_sol_1741737600
+2025-04-29 01:57:03,015 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Starting training SOL/USDT --------------------
+2025-04-29 01:57:03,042 - freqtrade.freqai.data_kitchen - INFO - SOL/USDT: dropped 0 training points due to NaNs in populated dataset 14400.
+2025-04-29 01:57:03,042 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Training on data from 2025-02-10 to 2025-03-11 --------------------
+2025-04-29 01:57:08,138 - datasieve.pipeline - INFO - DI tossed 3 predictions for being too far from training data.
+2025-04-29 01:57:08,141 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 111 features
+2025-04-29 01:57:08,141 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 11520 data points
+2025-04-29 01:57:09,614 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Done training SOL/USDT (6.60 secs) --------------------
+2025-04-29 01:57:09,615 - freqtrade.freqai.freqai_interface - INFO - Saving metadata to disk.
+2025-04-29 01:57:10,150 - freqtrade.freqai.freqai_interface - INFO - Training SOL/USDT, 2/2 pairs from 2025-02-20 00:00:00 to 2025-03-22 00:00:00, 9/11 trains
+2025-04-29 01:57:10,151 - freqtrade.freqai.data_kitchen - INFO - Could not find backtesting prediction file at 
+/freqtrade/user_data/models/test175/backtesting_predictions/cb_sol_1742601600_prediction.feather
+2025-04-29 01:57:10,159 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：SOL/USDT
+2025-04-29 01:57:10,167 - FreqaiExampleStrategy - INFO - 目标列形状：(52850,)
+2025-04-29 01:57:10,168 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.002704  49.623338
+1    0.003044  49.623338
+2    0.000465  49.623338
+3   -0.000380  49.623338
+4    0.002829  49.623338
+2025-04-29 01:57:10,181 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：SOL/USDT
+2025-04-29 01:57:10,188 - FreqaiExampleStrategy - INFO - 目标列形状：(57650,)
+2025-04-29 01:57:10,190 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.002704  49.644115
+1    0.003044  49.644115
+2    0.000465  49.644115
+3   -0.000380  49.644115
+4    0.002829  49.644115
+2025-04-29 01:57:10,195 - freqtrade.freqai.freqai_interface - INFO - Could not find model at /freqtrade/user_data/models/test175/sub-train-SOL_1742601600/cb_sol_1742601600
+2025-04-29 01:57:10,196 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Starting training SOL/USDT --------------------
+2025-04-29 01:57:10,218 - freqtrade.freqai.data_kitchen - INFO - SOL/USDT: dropped 0 training points due to NaNs in populated dataset 14400.
+2025-04-29 01:57:10,218 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Training on data from 2025-02-20 to 2025-03-21 --------------------
+2025-04-29 01:57:15,185 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 111 features
+2025-04-29 01:57:15,186 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 11520 data points
+2025-04-29 01:57:16,538 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Done training SOL/USDT (6.34 secs) --------------------
+2025-04-29 01:57:16,539 - freqtrade.freqai.freqai_interface - INFO - Saving metadata to disk.
+2025-04-29 01:57:17,068 - freqtrade.freqai.freqai_interface - INFO - Training SOL/USDT, 2/2 pairs from 2025-03-02 00:00:00 to 2025-04-01 00:00:00, 10/11 trains
+2025-04-29 01:57:17,069 - freqtrade.freqai.data_kitchen - INFO - Could not find backtesting prediction file at 
+/freqtrade/user_data/models/test175/backtesting_predictions/cb_sol_1743465600_prediction.feather
+2025-04-29 01:57:17,084 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：SOL/USDT
+2025-04-29 01:57:17,092 - FreqaiExampleStrategy - INFO - 目标列形状：(57650,)
+2025-04-29 01:57:17,094 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.002704  49.644115
+1    0.003044  49.644115
+2    0.000465  49.644115
+3   -0.000380  49.644115
+4    0.002829  49.644115
+2025-04-29 01:57:17,108 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：SOL/USDT
+2025-04-29 01:57:17,115 - FreqaiExampleStrategy - INFO - 目标列形状：(62450,)
+2025-04-29 01:57:17,117 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.002704  49.601082
+1    0.003044  49.601082
+2    0.000465  49.601082
+3   -0.000380  49.601082
+4    0.002829  49.601082
+2025-04-29 01:57:17,124 - freqtrade.freqai.freqai_interface - INFO - Could not find model at /freqtrade/user_data/models/test175/sub-train-SOL_1743465600/cb_sol_1743465600
+2025-04-29 01:57:17,125 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Starting training SOL/USDT --------------------
+2025-04-29 01:57:17,151 - freqtrade.freqai.data_kitchen - INFO - SOL/USDT: dropped 0 training points due to NaNs in populated dataset 14400.
+2025-04-29 01:57:17,151 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Training on data from 2025-03-02 to 2025-03-31 --------------------
+2025-04-29 01:57:22,430 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 111 features
+2025-04-29 01:57:22,430 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 11520 data points
+2025-04-29 01:57:23,725 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Done training SOL/USDT (6.60 secs) --------------------
+2025-04-29 01:57:23,726 - freqtrade.freqai.freqai_interface - INFO - Saving metadata to disk.
+2025-04-29 01:57:24,305 - freqtrade.freqai.freqai_interface - INFO - Training SOL/USDT, 2/2 pairs from 2025-03-12 00:00:00 to 2025-04-11 00:00:00, 11/11 trains
+2025-04-29 01:57:24,305 - freqtrade.freqai.data_kitchen - INFO - Could not find backtesting prediction file at 
+/freqtrade/user_data/models/test175/backtesting_predictions/cb_sol_1744329600_prediction.feather
+2025-04-29 01:57:24,318 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：SOL/USDT
+2025-04-29 01:57:24,325 - FreqaiExampleStrategy - INFO - 目标列形状：(62450,)
+2025-04-29 01:57:24,327 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.002704  49.601082
+1    0.003044  49.601082
+2    0.000465  49.601082
+3   -0.000380  49.601082
+4    0.002829  49.601082
+2025-04-29 01:57:24,337 - FreqaiExampleStrategy - INFO - 设置 FreqAI 目标，交易对：SOL/USDT
+2025-04-29 01:57:24,345 - FreqaiExampleStrategy - INFO - 目标列形状：(66770,)
+2025-04-29 01:57:24,346 - FreqaiExampleStrategy - INFO - 目标列预览：
+   up_or_down  &-buy_rsi
+0    0.002704  49.729824
+1    0.003044  49.729824
+2    0.000465  49.729824
+3   -0.000380  49.729824
+4    0.002829  49.729824
+2025-04-29 01:57:24,352 - freqtrade.freqai.freqai_interface - INFO - Could not find model at /freqtrade/user_data/models/test175/sub-train-SOL_1744329600/cb_sol_1744329600
+2025-04-29 01:57:24,353 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Starting training SOL/USDT --------------------
+2025-04-29 01:57:24,376 - freqtrade.freqai.data_kitchen - INFO - SOL/USDT: dropped 0 training points due to NaNs in populated dataset 14400.
+2025-04-29 01:57:24,376 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Training on data from 2025-03-12 to 2025-04-10 --------------------
+2025-04-29 01:57:29,392 - datasieve.pipeline - INFO - DI tossed 1948 predictions for being too far from training data.
+2025-04-29 01:57:29,396 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 111 features
+2025-04-29 01:57:29,396 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - Training model on 11520 data points
+2025-04-29 01:57:30,474 - freqtrade.freqai.base_models.BaseRegressionModel - INFO - -------------------- Done training SOL/USDT (6.12 secs) --------------------
+2025-04-29 01:57:30,475 - freqtrade.freqai.freqai_interface - INFO - Saving metadata to disk.
+2025-04-29 01:57:31,077 - FreqaiExampleStrategy - INFO - 动态参数：buy_rsi=50.0, sell_rsi=70.0, stoploss=-0.15, trailing_stop_positive=0.05
+2025-04-29 01:57:31,096 - FreqaiExampleStrategy - INFO - up_or_down 值统计：
+up_or_down
+0    33825
+1    32946
+2025-04-29 01:57:31,097 - FreqaiExampleStrategy - INFO - do_predict 值统计：
+do_predict
+0.0    36730
+1.0    30041
+2025-04-29 01:57:31,105 - freqtrade.optimize.backtesting - INFO - Backtesting with data from 2025-01-01 00:00:00 up to 2025-04-20 00:00:00 (109 days).
+2025-04-29 01:57:31,109 - FreqaiExampleStrategy - ERROR - MACD 或 MACD 信号列缺失，无法生成买入信号。尝试重新计算 MACD 列。
+2025-04-29 01:57:31,111 - FreqaiExampleStrategy - INFO - MACD 列已成功重新计算。
+2025-04-29 01:57:31,193 - FreqaiExampleStrategy - ERROR - MACD 或 MACD 信号列缺失，无法生成买入信号。尝试重新计算 MACD 列。
+2025-04-29 01:57:31,195 - FreqaiExampleStrategy - INFO - MACD 列已成功重新计算。
+2025-04-29 01:57:33,776 - freqtrade.misc - INFO - dumping json to "/freqtrade/user_data/backtest_results/backtest-result-2025-04-29_01-57-33.meta.json"
+Result for strategy FreqaiExampleStrategy
+                                             BACKTESTING REPORT                                              
+┏━━━━━━━━━━┳━━━━━━━━┳━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┓
+┃     Pair ┃ Trades ┃ Avg Profit % ┃ Tot Profit USDT ┃ Tot Profit % ┃ Avg Duration ┃  Win  Draw  Loss  Win% ┃
+┡━━━━━━━━━━╇━━━━━━━━╇━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━┩
+│ BTC/USDT │     38 │        -0.39 │         -22.029 │         -2.2 │     22:13:00 │    5    32     1  13.2 │
+│ SOL/USDT │     44 │        -1.94 │        -128.236 │       -12.82 │     16:35:00 │   12    26     6  27.3 │
+│    TOTAL │     82 │        -1.22 │        -150.265 │       -15.03 │     19:12:00 │   17    58     7  20.7 │
+└──────────┴────────┴──────────────┴─────────────────┴──────────────┴──────────────┴────────────────────────┘
+                                         LEFT OPEN TRADES REPORT                                          
+┏━━━━━━━┳━━━━━━━━┳━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┓
+┃  Pair ┃ Trades ┃ Avg Profit % ┃ Tot Profit USDT ┃ Tot Profit % ┃ Avg Duration ┃  Win  Draw  Loss  Win% ┃
+┡━━━━━━━╇━━━━━━━━╇━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━┩
+│ TOTAL │      0 │          0.0 │           0.000 │          0.0 │         0:00 │    0     0     0     0 │
+└───────┴────────┴──────────────┴─────────────────┴──────────────┴──────────────┴────────────────────────┘
+                                                ENTER TAG STATS                                                
+┏━━━━━━━━━━━┳━━━━━━━━━┳━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┓
+┃ Enter Tag ┃ Entries ┃ Avg Profit % ┃ Tot Profit USDT ┃ Tot Profit % ┃ Avg Duration ┃  Win  Draw  Loss  Win% ┃
+┡━━━━━━━━━━━╇━━━━━━━━━╇━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━┩
+│      long │      82 │        -1.22 │        -150.265 │       -15.03 │     19:12:00 │   17    58     7  20.7 │
+│     TOTAL │      82 │        -1.22 │        -150.265 │       -15.03 │     19:12:00 │   17    58     7  20.7 │
+└───────────┴─────────┴──────────────┴─────────────────┴──────────────┴──────────────┴────────────────────────┘
+                                                    EXIT REASON STATS                                                     
+┏━━━━━━━━━━━━━━━━━━━━┳━━━━━━━┳━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┓
+┃        Exit Reason ┃ Exits ┃ Avg Profit % ┃ Tot Profit USDT ┃ Tot Profit % ┃     Avg Duration ┃  Win  Draw  Loss  Win% ┃
+┡━━━━━━━━━━━━━━━━━━━━╇━━━━━━━╇━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━┩
+│                roi │    75 │         0.07 │           7.926 │         0.79 │         14:48:00 │   17    58     0   100 │
+│ trailing_stop_loss │     7 │       -15.04 │        -158.191 │       -15.82 │ 2 days, 18:13:00 │    0     0     7     0 │
+│              TOTAL │    82 │        -1.22 │        -150.265 │       -15.03 │         19:12:00 │   17    58     7  20.7 │
+└────────────────────┴───────┴──────────────┴─────────────────┴──────────────┴──────────────────┴────────────────────────┘
+                                                                   MIXED TAG STATS                                                                   
+┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━┳━━━━━━━━┳━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┓
+┃                      Enter Tag ┃ Exit Reason ┃ Trades ┃ Avg Profit % ┃ Tot Profit USDT ┃ Tot Profit % ┃     Avg Duration ┃  Win  Draw  Loss  Win% ┃
+┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━╇━━━━━━━━╇━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━┩
+│                ('long', 'roi') │             │     75 │         0.07 │           7.926 │         0.79 │         14:48:00 │   17    58     0   100 │
+│ ('long', 'trailing_stop_loss') │             │      7 │       -15.04 │        -158.191 │       -15.82 │ 2 days, 18:13:00 │    0     0     7     0 │
+│                          TOTAL │             │     82 │        -1.22 │        -150.265 │       -15.03 │         19:12:00 │   17    58     7  20.7 │
+└────────────────────────────────┴─────────────┴────────┴──────────────┴─────────────────┴──────────────┴──────────────────┴────────────────────────┘
+                   SUMMARY METRICS                   
+┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━┓
+┃ Metric                      ┃ Value               ┃
+┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━┩
+│ Backtesting from            │ 2025-01-01 00:00:00 │
+│ Backtesting to              │ 2025-04-20 00:00:00 │
+│ Trading Mode                │ Spot                │
+│ Max open trades             │ 2                   │
+│                             │                     │
+│ Total/Daily Avg Trades      │ 82 / 0.75           │
+│ Starting balance            │ 1000 USDT           │
+│ Final balance               │ 849.735 USDT        │
+│ Absolute profit             │ -150.265 USDT       │
+│ Total profit %              │ -15.03%             │
+│ CAGR %                      │ -42.03%             │
+│ Sortino                     │ -252.56             │
+│ Sharpe                      │ -4.15               │
+│ Calmar                      │ -17.48              │
+│ SQN                         │ -2.60               │
+│ Profit factor               │ 0.05                │
+│ Expectancy (Ratio)          │ -1.83 (-0.79)       │
+│ Avg. daily profit %         │ -0.14%              │
+│ Avg. stake amount           │ 150 USDT            │
+│ Total trade volume          │ 24523.15 USDT       │
+│                             │                     │
+│ Best Pair                   │ BTC/USDT -2.20%     │
+│ Worst Pair                  │ SOL/USDT -12.82%    │
+│ Best trade                  │ SOL/USDT 0.90%      │
+│ Worst trade                 │ SOL/USDT -15.19%    │
+│ Best day                    │ 1.76 USDT           │
+│ Worst day                   │ -22.827 USDT        │
+│ Days win/draw/lose          │ 14 / 80 / 7         │
+│ Avg. Duration Winners       │ 0:55:00             │
+│ Avg. Duration Loser         │ 2 days, 18:13:00    │
+│ Max Consecutive Wins / Loss │ 2 / 16              │
+│ Rejected Entry signals      │ 0                   │
+│ Entry/Exit Timeouts         │ 0 / 0               │
+│                             │                     │
+│ Min balance                 │ 849.735 USDT        │
+│ Max balance                 │ 1000.508 USDT       │
+│ Max % of account underwater │ 15.07%              │
+│ Absolute Drawdown (Account) │ 15.07%              │
+│ Absolute Drawdown           │ 150.773 USDT        │
+│ Drawdown high               │ 0.508 USDT          │
+│ Drawdown low                │ -150.265 USDT       │
+│ Drawdown Start              │ 2025-01-06 19:48:00 │
+│ Drawdown End                │ 2025-04-06 23:15:00 │
+│ Market change               │ -26.79%             │
+└─────────────────────────────┴─────────────────────┘
+
+Backtested 2025-01-01 00:00:00 -> 2025-04-20 00:00:00 | Max open trades : 2
+                                                                STRATEGY SUMMARY                                                                 
+┏━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━┳━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━┓
+┃              Strategy ┃ Trades ┃ Avg Profit % ┃ Tot Profit USDT ┃ Tot Profit % ┃ Avg Duration ┃  Win  Draw  Loss  Win% ┃             Drawdown ┃
+┡━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━╇━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━┩
+│ FreqaiExampleStrategy │     82 │        -1.22 │        -150.265 │       -15.03 │     19:12:00 │   17    58     7  20.7 │ 150.773 USDT  15.07% │
+└───────────────────────┴────────┴──────────────┴─────────────────┴──────────────┴──────────────┴────────────────────────┴──────────────────────┘

freqtrade/templates/FreqaiExampleHybridStrategy.py

@@ -1,311 +0,0 @@
-import logging
-
-import numpy as np  # noqa
-import pandas as pd  # noqa
-import talib.abstract as ta
-from pandas import DataFrame
-from technical import qtpylib
-
-from freqtrade.strategy import IntParameter, IStrategy, merge_informative_pair 
-
-
-logger = logging.getLogger(__name__)
-
-
-class FreqaiExampleHybridStrategy(IStrategy):
-    """
-    Example of a hybrid FreqAI strat, designed to illustrate how a user may employ
-    FreqAI to bolster a typical Freqtrade strategy.
-
-    Launching this strategy would be:
-
-    freqtrade trade --strategy FreqaiExampleHybridStrategy --strategy-path freqtrade/templates
-    --freqaimodel CatboostClassifier --config config_examples/config_freqai.example.json
-
-    or the user simply adds this to their config:
-
-    "freqai": {
-        "enabled": true,
-        "purge_old_models": 2,
-        "train_period_days": 15,
-        "identifier": "unique-id",
-        "feature_parameters": {
-            "include_timeframes": [
-                "3m",
-                "15m",
-                "1h"
-            ],
-            "include_corr_pairlist": [
-                "BTC/USDT",
-                "ETH/USDT"
-            ],
-            "label_period_candles": 20,
-            "include_shifted_candles": 2,
-            "DI_threshold": 0.9,
-            "weight_factor": 0.9,
-            "principal_component_analysis": false,
-            "use_SVM_to_remove_outliers": true,
-            "indicator_periods_candles": [10, 20]
-        },
-        "data_split_parameters": {
-            "test_size": 0,
-            "random_state": 1
-        },
-        "model_training_parameters": {
-            "n_estimators": 800
-        }
-    },
-
-    Thanks to @smarmau and @johanvulgt for developing and sharing the strategy.
-    """
-
-    minimal_roi = {
-        # "120": 0.0,  # exit after 120 minutes at break even
-        "60": 0.01,
-        "30": 0.02,
-        "0": 0.04,
-    }
-
-    plot_config = {
-        "main_plot": {
-            "tema": {},
-        },
-        "subplots": {
-            "MACD": {
-                "macd": {"color": "blue"},
-                "macdsignal": {"color": "orange"},
-            },
-            "RSI": {
-                "rsi": {"color": "red"},
-            },
-            "Up_or_down": {
-                "&s-up_or_down": {"color": "green"},
-            },
-        },
-    }
-
-    process_only_new_candles = True
-    stoploss = -0.05
-    use_exit_signal = True
-    startup_candle_count: int = 30
-    can_short = False
-
-    # Hyperoptable parameters
-    buy_rsi = IntParameter(low=1, high=50, default=30, space="buy", optimize=True, load=True)
-    sell_rsi = IntParameter(low=50, high=100, default=70, space="sell", optimize=True, load=True)
-
-    def feature_engineering_expand_all(
-        self, dataframe: DataFrame, period: int, metadata: dict, **kwargs
-    ) -> DataFrame:
-        """
-        *Only functional with FreqAI enabled strategies*
-        This function will automatically expand the defined features on the config defined
-        `indicator_periods_candles`, `include_timeframes`, `include_shifted_candles`, and
-        `include_corr_pairs`. In other words, a single feature defined in this function
-        will automatically expand to a total of
-        `indicator_periods_candles` * `include_timeframes` * `include_shifted_candles` *
-        `include_corr_pairs` numbers of features added to the model.
-
-        All features must be prepended with `%` to be recognized by FreqAI internals.
-
-        More details on how these config defined parameters accelerate feature engineering
-        in the documentation at:
-
-        https://www.freqtrade.io/en/latest/freqai-parameter-table/#feature-parameters
-
-        https://www.freqtrade.io/en/latest/freqai-feature-engineering/#defining-the-features
-
-        :param dataframe: strategy dataframe which will receive the features
-        :param period: period of the indicator - usage example:
-        :param metadata: metadata of current pair
-        dataframe["%-ema-period"] = ta.EMA(dataframe, timeperiod=period)
-        """
-
-        dataframe["%-rsi-period"] = ta.RSI(dataframe, timeperiod=period)
-        dataframe["%-mfi-period"] = ta.MFI(dataframe, timeperiod=period)
-        dataframe["%-adx-period"] = ta.ADX(dataframe, timeperiod=period)
-        dataframe["%-sma-period"] = ta.SMA(dataframe, timeperiod=period)
-        dataframe["%-ema-period"] = ta.EMA(dataframe, timeperiod=period)
-
-        bollinger = qtpylib.bollinger_bands(
-            qtpylib.typical_price(dataframe), window=period, stds=2.2
-        )
-        dataframe["bb_lowerband-period"] = bollinger["lower"]
-        dataframe["bb_middleband-period"] = bollinger["mid"]
-        dataframe["bb_upperband-period"] = bollinger["upper"]
-
-        dataframe["%-bb_width-period"] = (
-            dataframe["bb_upperband-period"] - dataframe["bb_lowerband-period"]
-        ) / dataframe["bb_middleband-period"]
-        dataframe["%-close-bb_lower-period"] = dataframe["close"] / dataframe["bb_lowerband-period"]
-
-        dataframe["%-roc-period"] = ta.ROC(dataframe, timeperiod=period)
-
-        dataframe["%-relative_volume-period"] = (
-            dataframe["volume"] / dataframe["volume"].rolling(period).mean()
-        )
-
-        return dataframe
-
-    def feature_engineering_expand_basic(
-        self, dataframe: DataFrame, metadata: dict, **kwargs
-    ) -> DataFrame:
-        """
-        *Only functional with FreqAI enabled strategies*
-        This function will automatically expand the defined features on the config defined
-        `include_timeframes`, `include_shifted_candles`, and `include_corr_pairs`.
-        In other words, a single feature defined in this function
-        will automatically expand to a total of
-        `include_timeframes` * `include_shifted_candles` * `include_corr_pairs`
-        numbers of features added to the model.
-
-        Features defined here will *not* be automatically duplicated on user defined
-        `indicator_periods_candles`
-
-        All features must be prepended with `%` to be recognized by FreqAI internals.
-
-        More details on how these config defined parameters accelerate feature engineering
-        in the documentation at:
-
-        https://www.freqtrade.io/en/latest/freqai-parameter-table/#feature-parameters
-
-        https://www.freqtrade.io/en/latest/freqai-feature-engineering/#defining-the-features
-
-        :param dataframe: strategy dataframe which will receive the features
-        :param metadata: metadata of current pair
-        dataframe["%-pct-change"] = dataframe["close"].pct_change()
-        dataframe["%-ema-200"] = ta.EMA(dataframe, timeperiod=200)
-        """
-        dataframe["%-pct-change"] = dataframe["close"].pct_change()
-        dataframe["%-raw_volume"] = dataframe["volume"]
-        dataframe["%-raw_price"] = dataframe["close"]
-        return dataframe
-
-    def feature_engineering_standard(
-        self, dataframe: DataFrame, metadata: dict, **kwargs
-    ) -> DataFrame:
-        """
-        *Only functional with FreqAI enabled strategies*
-        This optional function will be called once with the dataframe of the base timeframe.
-        This is the final function to be called, which means that the dataframe entering this
-        function will contain all the features and columns created by all other
-        freqai_feature_engineering_* functions.
-
-        This function is a good place to do custom exotic feature extractions (e.g. tsfresh).
-        This function is a good place for any feature that should not be auto-expanded upon
-        (e.g. day of the week).
-
-        All features must be prepended with `%` to be recognized by FreqAI internals.
-
-        More details about feature engineering available:
-
-        https://www.freqtrade.io/en/latest/freqai-feature-engineering
-
-        :param dataframe: strategy dataframe which will receive the features
-        :param metadata: metadata of current pair
-        usage example: dataframe["%-day_of_week"] = (dataframe["date"].dt.dayofweek + 1) / 7
-        """
-        dataframe["%-day_of_week"] = dataframe["date"].dt.dayofweek
-        dataframe["%-hour_of_day"] = dataframe["date"].dt.hour
-        return dataframe
-    def set_freqai_targets(self, dataframe: DataFrame, metadata: dict, **kwargs) -> DataFrame:
-        logger.info(f"Setting FreqAI targets for pair: {metadata['pair']}")
-        logger.info(f"DataFrame shape: {dataframe.shape}")
-        logger.info(f"Available columns: {list(dataframe.columns)}")
-        logger.info(f"First few rows:\n{dataframe[['date', 'close']].head().to_string()}")
-        
-        if "close" not in dataframe.columns:
-            logger.error("Required 'close' column missing in dataframe")
-            raise ValueError("Required 'close' column missing in dataframe")
-        
-        if len(dataframe) < 50:
-            logger.error(f"Insufficient data: {len(dataframe)} rows, need at least 50 for shift(-50)")
-            raise ValueError("Insufficient data for target calculation")
-        
-        try:
-            # 生成数值型标签：1 表示上涨，0 表示下跌
-            dataframe["&-up_or_down"] = np.where(
-                dataframe["close"].shift(-50) > dataframe["close"],
-                1.0,  # 数值型标签
-                0.0
-            )
-        except Exception as e:
-            logger.error(f"Failed to create &-up_or_down column: {str(e)}")
-            raise
-        
-        logger.info(f"Target column head:\n{dataframe[['&-up_or_down']].head().to_string()}")
-        
-        if "&-up_or_down" not in dataframe.columns:
-            logger.error("FreqAI failed to generate the &-up_or_down column")
-            raise KeyError("FreqAI failed to generate the &-up_or_down column")
-        
-        logger.info("FreqAI targets set successfully")
-        return dataframe
-
-    def populate_indicators(self, dataframe: DataFrame, metadata: dict) -> DataFrame:
-        logger.info(f"Processing pair: {metadata['pair']}")
-        logger.info(f"Input DataFrame shape: {dataframe.shape}")
-        logger.info(f"Input DataFrame columns: {list(dataframe.columns)}")
-        logger.info(f"Input DataFrame head:\n{dataframe[['date', 'close', 'volume']].head().to_string()}")
-        
-        # Ensure FreqAI processing
-        logger.info("Calling self.freqai.start")
-        try:
-            dataframe = self.freqai.start(dataframe, metadata, self)
-        except Exception as e:
-            logger.error(f"self.freqai.start failed: {str(e)}")
-            raise
-        logger.info("self.freqai.start completed")
-        
-        logger.info(f"Output DataFrame shape: {dataframe.shape}")
-        logger.info(f"Output DataFrame columns: {list(dataframe.columns)}")
-        # Safely log columns that exist
-        available_columns = [col for col in ['date', 'close', '&-up_or_down'] if col in dataframe.columns]
-        logger.info(f"Output DataFrame head:\n{dataframe[available_columns].head().to_string()}")
-        
-        if "&-up_or_down" not in dataframe.columns:
-            logger.error("FreqAI did not generate the required &-up_or_down column")
-            raise KeyError("FreqAI did not generate the required &-up_or_down column")
-
-        # RSI
-        dataframe["rsi"] = ta.RSI(dataframe)
-
-        # Bollinger Bands
-        bollinger = qtpylib.bollinger_bands(qtpylib.typical_price(dataframe), window=20, stds=2)
-        dataframe["bb_lowerband"] = bollinger["lower"]
-        dataframe["bb_middleband"] = bollinger["mid"]
-        dataframe["bb_upperband"] = bollinger["upper"]
-        dataframe["bb_percent"] = (dataframe["close"] - dataframe["bb_lowerband"]) / (
-            dataframe["bb_upperband"] - dataframe["bb_lowerband"]
-        )
-        dataframe["bb_width"] = (dataframe["bb_upperband"] - dataframe["bb_lowerband"]) / dataframe[
-            "bb_middleband"
-        ]
-
-        # TEMA
-        dataframe["tema"] = ta.TEMA(dataframe, timeperiod=9)
-
-        return dataframe
-    def populate_entry_trend(self, df: DataFrame, metadata: dict) -> DataFrame:
-        df.loc[
-            (
-                (qtpylib.crossed_above(df["rsi"], self.buy_rsi.value))
-                & (df["tema"] <= df["bb_middleband"])
-                & (df["tema"] > df["tema"].shift(1))
-                & (df["volume"] > 0)
-            ),
-            "enter_long",
-        ] = 1
-        return df
-
-    def populate_exit_trend(self, df: DataFrame, metadata: dict) -> DataFrame:
-        df.loc[
-            (
-                (qtpylib.crossed_above(df["rsi"], self.sell_rsi.value))
-                & (df["tema"] > df["bb_middleband"])
-                & (df["tema"] < df["tema"].shift(1))
-                & (df["volume"] > 0)
-            ),
-            "exit_long",
-        ] = 1
-        return df 

freqtrade/templates/FreqaiExampleStrategy.json

@@ -1,32 +0,0 @@
-{
-  "strategy_name": "FreqaiExampleStrategy",
-  "params": {
-    "trailing": {
-      "trailing_stop": true,
-      "trailing_stop_positive": 0.01,
-      "trailing_stop_positive_offset": 0.02,
-      "trailing_only_offset_is_reached": false
-    },
-    "max_open_trades": {
-      "max_open_trades": 4
-    },
-    "buy": {
-      "buy_rsi": 39.92672300850069
-    },
-    "sell": {
-      "sell_rsi": 69.92672300850067
-    },
-    "protection": {},
-    "roi": {
-      "0": 0.132,
-      "8": 0.047,
-      "14": 0.007,
-      "60": 0
-    },
-    "stoploss": {
-      "stoploss": -0.322
-    }
-  },
-  "ft_stratparam_v": 1,
-  "export_time": "2025-04-23 12:30:05.550433+00:00"
-}

freqtrade/templates/aa.json

@@ -0,0 +1,44 @@
+{
+  "strategy_name": "FreqaiPrimer",
+  "params": {
+    "stoploss": {
+      "stoploss": -0.05
+    },
+    "max_open_trades": {
+      "max_open_trades": 5
+    },
+    "buy": {
+      "ADD_POSITION_THRESHOLD": -0.046,
+      "BUY_THRESHOLD_MAX": -0.01,
+      "BUY_THRESHOLD_MIN": -0.026,
+      "COOLDOWN_PERIOD_MINUTES": 9,
+      "MAX_ENTRY_POSITION_ADJUSTMENT": 3
+    },
+    "sell": {
+      "EXIT_POSITION_RATIO": 0.572,
+      "HOLD_TIME_EXTENSION": 1.975,
+      "HOLD_TIME_REDUCTION": 0.832,
+      "MAX_HOLD_TIME_MINUTES": 101,
+      "SELL_THRESHOLD_MAX": 0.085,
+      "SELL_THRESHOLD_MIN": 0.007,
+      "SLOPE_THRESHOLD": 0.0,
+      "TRAILING_STOP_DISTANCE": 0.005,
+      "TRAILING_STOP_START": 0.048
+    },
+    "protection": {},
+    "roi": {
+      "0": 0.128,
+      "12": 0.038,
+      "36": 0.023,
+      "53": 0
+    },
+    "trailing": {
+      "trailing_stop": true,
+      "trailing_stop_positive": 0.09,
+      "trailing_stop_positive_offset": 0.16499999999999998,
+      "trailing_only_offset_is_reached": false
+    }
+  },
+  "ft_stratparam_v": 1,
+  "export_time": "2025-06-27 03:30:20.329593+00:00"
+}

freqtrade/templates/aaa.diff

@@ -0,0 +1,67 @@
+--- original_strategy.py
++++ modified_strategy.py
+@@ -28,8 +28,8 @@ class FreqaiPrimer(IStrategy):
+     TRAILING_STOP_DISTANCE = DecimalParameter(0.005, 0.02, default=0.01, space='sell', optimize=True)
+
+-    BUY_THRESHOLD_MIN = DecimalParameter(-0.1, -0.01, default=-0.05, space='buy', optimize=True)
+-    BUY_THRESHOLD_MAX = DecimalParameter(-0.02, -0.001, default=-0.005, space='buy', optimize=True)
++    BUY_THRESHOLD_MIN = DecimalParameter(-0.15, -0.03, default=-0.07, space='buy', optimize=True)
++    BUY_THRESHOLD_MAX = DecimalParameter(-0.04, -0.005, default=-0.01, space='buy', optimize=True)
+     SELL_THRESHOLD_MIN = DecimalParameter(0.001, 0.02, default=0.005, space='sell', optimize=True)
+     SELL_THRESHOLD_MAX = DecimalParameter(0.02, 0.1, default=0.05, space='sell', optimize=True)
+
+@@ -356,10 +356,10 @@ class FreqaiPrimer(IStrategy):
+         volume_z_score_min = 0.5
+         volume_z_score_max = 1.5
+         volume_z_score_threshold = self.linear_map(trend_score, 0, 100, volume_z_score_max, volume_z_score_min)
+
+-        rsi_min = 40
+-        rsi_max = 60
++        rsi_min = 35
++        rsi_max = 55
+         rsi_threshold = self.linear_map(trend_score, 0, 100, rsi_max, rsi_min)
+
+-        stochrsi_min = 30
+-        stochrsi_max = 50
++        stochrsi_min = 25
++        stochrsi_max = 45
+         stochrsi_threshold = self.linear_map(trend_score, 0, 100, stochrsi_max, stochrsi_min)
+
+         if "&-price_value_divergence" in dataframe.columns:
+             # 计算 STOCHRSI
+             stochrsi = ta.STOCHRSI(dataframe, timeperiod=14, fastk_period=3, fastd_period=3)
+             dataframe["stochrsi_k"] = stochrsi["fastk"]
++            # 添加 15m 时间框架的 EMA200
++            dataframe_15m = self.dp.get_pair_dataframe(pair=pair, timeframe="15m")
++            dataframe_15m["ema200_15m"] = ta.EMA(dataframe_15m, timeperiod=200)
++            dataframe["ema200_15m"] = dataframe_15m["ema200_15m"].reindex(dataframe.index, method="ffill")
+
+             cond1 = (dataframe["&-price_value_divergence"] < self.buy_threshold)
+             cond2 = (dataframe["volume_z_score"] > volume_z_score_threshold)
+             cond3 = (dataframe["rsi"] < rsi_threshold)
+             cond4 = (dataframe["close"] <= dataframe["bb_lowerband"])
+-            cond5 = (dataframe["stochrsi_k"] < stochrsi_threshold)
+-            buy_condition = cond1 & cond2 & cond3 & cond4 & cond5
++            cond5 = (dataframe["stochrsi_k"] < stochrsi_threshold)
++            cond6 = (dataframe["close"] < dataframe["ema200_15m"])  # 价格低于 15m EMA200
++            buy_condition = cond1 & cond2 & cond3 & cond4 & cond5 & cond6
+             conditions.append(buy_condition)
+
+@@ -534,7 +534,7 @@ class FreqaiPrimer(IStrategy):
+     def get_market_trend(self, dataframe: DataFrame = None, metadata: dict = None) -> int:
+         try:
+             timeframes = ["3m", "15m", "1h"]
+-            weights = {"3m": 0.5, "15m": 0.3, "1h": 0.2}
++            weights = {"3m": 0.3, "15m": 0.35, "1h": 0.35}
+             trend_scores = {}
+             pair = metadata.get('pair', 'Unknown') if metadata else 'Unknown'
+
+@@ -565,7 +565,7 @@ class FreqaiPrimer(IStrategy):
+                 pair_df["ema_short_slope"] = (pair_df["ema_short"] - pair_df["ema_short"].shift(10)) / pair_df["ema_short"].shift(10)
+
+             # 动态调整权重
+-            if trend_scores.get("1h", 50) - trend_scores.get("3m", 50) > 20:
+-                weights = {"3m": 0.3, "15m": 0.3, "1h": 0.4}
++            if trend_scores.get("1h", 50) - trend_scores.get("3m", 50) > 20 or trend_scores.get("15m", 50) - trend_scores.get("3m", 50) > 20:
++                weights = {"3m": 0.2, "15m": 0.35, "1h": 0.45}
+                 logger.debug(f"[{pair}] 1h 趋势得分（{trend_scores.get('1h', 50)}）显著高于 3m（{trend_scores.get('3m', 50)}），调整权重为 {weights}")

freqtrade/templates/base_config.json.j2

@@ -1,69 +0,0 @@
-{%set volume_pairlist = '{
-            "method": "VolumePairList",
-            "number_assets": 20,
-            "sort_key": "quoteVolume",
-            "min_value": 0,
-            "refresh_period": 1800
-        }' %}
-{
-    "$schema": "https://schema.freqtrade.io/schema.json",
-    "max_open_trades": {{ max_open_trades }},
-    "stake_currency": "{{ stake_currency }}",
-    "stake_amount": {{ stake_amount }},
-    "tradable_balance_ratio": 0.99,
-{{- ('\n    "fiat_display_currency": "' + fiat_display_currency + '",') if fiat_display_currency else ''}}
-{{- ('\n    "timeframe": "' + timeframe + '",') if timeframe else '' }}
-    "dry_run": {{ dry_run | lower }},
-    "dry_run_wallet": 1000,
-    "cancel_open_orders_on_exit": false,
-    "trading_mode": "{{ trading_mode }}",
-    "margin_mode": "{{ margin_mode }}",
-    "unfilledtimeout": {
-        "entry": 10,
-        "exit": 10,
-        "exit_timeout_count": 0,
-        "unit": "minutes"
-    },
-    "entry_pricing": {
-        "price_side": "same",
-        "use_order_book": true,
-        "order_book_top": 1,
-        "price_last_balance": 0.0,
-        "check_depth_of_market": {
-            "enabled": false,
-            "bids_to_ask_delta": 1
-        }
-    },
-    "exit_pricing":{
-        "price_side": "same",
-        "use_order_book": true,
-        "order_book_top": 1
-    },
-    {{ exchange | indent(4) }},
-    "pairlists": [
-        {{ volume_pairlist }}
-    ],
-    "telegram": {
-        "enabled": {{ telegram | lower }},
-        "token": "{{ telegram_token }}",
-        "chat_id": "{{ telegram_chat_id }}"
-    },
-    "api_server": {
-        "enabled": {{ api_server | lower }},
-        "listen_ip_address": "{{ api_server_listen_addr | default("127.0.0.1", true) }}",
-        "listen_port": 8080,
-        "verbosity": "error",
-        "enable_openapi": false,
-        "jwt_secret_key": "{{ api_server_jwt_key }}",
-        "ws_token": "{{ api_server_ws_token }}",
-        "CORS_origins": [],
-        "username": "{{ api_server_username }}",
-        "password": "{{ api_server_password }}"
-    },
-    "bot_name": "freqtrade",
-    "initial_state": "running",
-    "force_entry_enable": false,
-    "internals": {
-        "process_throttle_secs": 5
-    }
-}

freqtrade/templates/base_strategy.py.j2

@@ -1,179 +0,0 @@
-# pragma pylint: disable=missing-docstring, invalid-name, pointless-string-statement
-# flake8: noqa: F401
-# isort: skip_file
-# --- Do not remove these imports ---
-import numpy as np
-import pandas as pd
-from datetime import datetime, timedelta, timezone
-from pandas import DataFrame
-from typing import Dict, Optional, Union, Tuple
-
-from freqtrade.strategy import (
-    IStrategy,
-    Trade,
-    Order,
-    PairLocks,
-    informative,  # @informative decorator
-    # Hyperopt Parameters
-    BooleanParameter,
-    CategoricalParameter,
-    DecimalParameter,
-    IntParameter,
-    RealParameter,
-    # timeframe helpers
-    timeframe_to_minutes,
-    timeframe_to_next_date,
-    timeframe_to_prev_date,
-    # Strategy helper functions
-    merge_informative_pair,
-    stoploss_from_absolute,
-    stoploss_from_open,
-)
-
-# --------------------------------
-# Add your lib to import here
-import talib.abstract as ta
-import pandas_ta as pta
-from technical import qtpylib
-
-
-class {{ strategy }}(IStrategy):
-    """
-    This is a strategy template to get you started.
-    More information in https://www.freqtrade.io/en/latest/strategy-customization/
-
-    You can:
-        :return: a Dataframe with all mandatory indicators for the strategies
-    - Rename the class name (Do not forget to update class_name)
-    - Add any methods you want to build your strategy
-    - Add any lib you need to build your strategy
-
-    You must keep:
-    - the lib in the section "Do not remove these libs"
-    - the methods: populate_indicators, populate_entry_trend, populate_exit_trend
-    You should keep:
-    - timeframe, minimal_roi, stoploss, trailing_*
-    """
-    # Strategy interface version - allow new iterations of the strategy interface.
-    # Check the documentation or the Sample strategy to get the latest version.
-    INTERFACE_VERSION = 3
-
-    # Optimal timeframe for the strategy.
-    timeframe = "5m"
-
-    # Can this strategy go short?
-    can_short: bool = False
-
-    # Minimal ROI designed for the strategy.
-    # This attribute will be overridden if the config file contains "minimal_roi".
-    minimal_roi = {
-        "60": 0.01,
-        "30": 0.02,
-        "0": 0.04
-    }
-
-    # Optimal stoploss designed for the strategy.
-    # This attribute will be overridden if the config file contains "stoploss".
-    stoploss = -0.10
-
-    # Trailing stoploss
-    trailing_stop = False
-    # trailing_only_offset_is_reached = False
-    # trailing_stop_positive = 0.01
-    # trailing_stop_positive_offset = 0.0  # Disabled / not configured
-
-    # Run "populate_indicators()" only for new candle.
-    process_only_new_candles = True
-
-    # These values can be overridden in the config.
-    use_exit_signal = True
-    exit_profit_only = False
-    ignore_roi_if_entry_signal = False
-
-    # Number of candles the strategy requires before producing valid signals
-    startup_candle_count: int = 30
-
-    # Strategy parameters
-    buy_rsi = IntParameter(10, 40, default=30, space="buy")
-    sell_rsi = IntParameter(60, 90, default=70, space="sell")
-
-    {{- attributes | indent(4) }}
-    {{- plot_config | indent(4) }}
-
-    def informative_pairs(self):
-        """
-        Define additional, informative pair/interval combinations to be cached from the exchange.
-        These pair/interval combinations are non-tradeable, unless they are part
-        of the whitelist as well.
-        For more information, please consult the documentation
-        :return: List of tuples in the format (pair, interval)
-            Sample: return [("ETH/USDT", "5m"),
-                            ("BTC/USDT", "15m"),
-                            ]
-        """
-        return []
-
-    def populate_indicators(self, dataframe: DataFrame, metadata: dict) -> DataFrame:
-        """
-        Adds several different TA indicators to the given DataFrame
-
-        Performance Note: For the best performance be frugal on the number of indicators
-        you are using. Let uncomment only the indicator you are using in your strategies
-        or your hyperopt configuration, otherwise you will waste your memory and CPU usage.
-        :param dataframe: Dataframe with data from the exchange
-        :param metadata: Additional information, like the currently traded pair
-        :return: a Dataframe with all mandatory indicators for the strategies
-        """
-        {{- indicators | indent(8) }}
-
-        return dataframe
-
-    def populate_entry_trend(self, dataframe: DataFrame, metadata: dict) -> DataFrame:
-        """
-        Based on TA indicators, populates the entry signal for the given dataframe
-        :param dataframe: DataFrame
-        :param metadata: Additional information, like the currently traded pair
-        :return: DataFrame with entry columns populated
-        """
-        dataframe.loc[
-            (
-                {{ buy_trend | indent(16) }}
-                (dataframe["volume"] > 0)  # Make sure Volume is not 0
-            ),
-            "enter_long"] = 1
-        # Uncomment to use shorts (Only used in futures/margin mode. Check the documentation for more info)
-        """
-        dataframe.loc[
-            (
-                {{ sell_trend | indent(16) }}
-                (dataframe['volume'] > 0)  # Make sure Volume is not 0
-            ),
-            'enter_short'] = 1
-        """
-
-        return dataframe
-
-    def populate_exit_trend(self, dataframe: DataFrame, metadata: dict) -> DataFrame:
-        """
-        Based on TA indicators, populates the exit signal for the given dataframe
-        :param dataframe: DataFrame
-        :param metadata: Additional information, like the currently traded pair
-        :return: DataFrame with exit columns populated
-        """
-        dataframe.loc[
-            (
-                {{ sell_trend | indent(16) }}
-                (dataframe["volume"] > 0)  # Make sure Volume is not 0
-            ),
-            "exit_long"] = 1
-        # Uncomment to use shorts (Only used in futures/margin mode. Check the documentation for more info)
-        """
-        dataframe.loc[
-            (
-                {{ buy_trend | indent(16) }}
-                (dataframe['volume'] > 0)  # Make sure Volume is not 0
-            ),
-            'exit_short'] = 1
-        """
-        return dataframe
-    {{- additional_methods | indent(4) }}

freqtrade/templates/chat.md

@@ -0,0 +1,85 @@
+> 请把当前策略的运行逻辑进行梳理和讲解
+
+我来为您梳理和讲解这个策略的运行逻辑。这是一个基于 FreqAI 的机器学习交易策略，主要使用 RSI 指标和机器学习模型进行交易决策。以下是策略的主要逻辑：
+
+### 1. 策略配置
+```python
+minimal_roi = {
+    "0": 0.02,
+    "7": 0.01,
+    "13": 0.005,
+    "60": 0
+}
+stoploss = 0.0
+trailing_stop = True
+```
+- 设置了静态的 ROI 目标
+- 启用了追踪止损
+
+### 2. 特征工程
+策略通过多个方法构建特征：
+- `feature_engineering_expand_all`: 计算 RSI 和布林带宽度
+- `feature_engineering_expand_basic`: 计算价格变化百分比
+- `feature_engineering_standard`: 计算波动率和时间特征
+
+### 3. 目标变量设置
+在 `set_freqai_targets` 方法中：
+- 使用未来 `label_period_candles` 周期的 RSI 值作为目标变量
+- 计算并标准化波动率特征
+
+### 4. 指标计算
+`populate_indicators` 方法中计算了多个技术指标：
+- RSI
+- 波动率
+- TEMA（三重指数移动平均线）
+- 布林带下轨
+- 未来价格方向（up_or_down）
+
+### 5. 机器学习模型
+使用 LightGBM 回归模型：
+```python
+freqai_info = {
+    "model": "LightGBMRegressor",
+    "feature_parameters": {...},
+    "data_split_parameters": {...},
+    "model_training_parameters": {...}
+}
+```
+- 模型预测买入和卖出的 RSI 值
+- 使用 SOL/USDT 作为相关交易对
+
+### 6. 入场逻辑
+`populate_entry_trend` 方法中：
+- 当实际 RSI 上穿预测的买入 RSI 时
+- TEMA 指标向上
+- 交易量大于 0
+- 模型预测结果为 1
+- 未来价格方向为上涨
+
+### 7. 出场逻辑
+`populate_exit_trend` 方法中：
+- 当实际 RSI 上穿预测的卖出 RSI 时
+- 或价格下跌超过 2%
+- 或价格跌破布林带下轨
+- 或持有时间超过 1 天
+
+### 8. 交易确认
+`confirm_trade_entry` 方法：
+- 检查当前价格是否在可接受范围内
+- 记录交易确认信息
+
+### 策略特点
+1. 机器学习驱动：使用 LightGBM 模型预测 RSI 值
+2. 多特征融合：结合技术指标、波动率和时间特征
+3. 风险控制：设置了 ROI 目标和追踪止损
+4. 动态调整：通过机器学习模型动态调整交易信号
+
+### 运行流程
+1. 数据准备：获取历史数据，计算技术指标
+2. 特征工程：构建机器学习特征
+3. 模型训练：使用历史数据训练 LightGBM 模型
+4. 信号生成：根据模型预测结果生成交易信号
+5. 交易执行：根据入场和出场逻辑执行交易
+6. 风险控制：根据 ROI 和止损规则管理仓位
+
+这个策略结合了传统技术分析和机器学习方法，通过模型预测来增强交易信号的准确性，同时保持了严格的风险控制机制。

freqtrade/templates/df.diff

@@ -0,0 +1,64 @@
+--- original_strategy.py
++++ modified_strategy.py
+@@ -297,6 +297,9 @@ class FreqaiPrimer(IStrategy):
+         dataframe["bb_lowerband"] = lowerband
+         dataframe["rsi"] = ta.RSI(dataframe, timeperiod=14)
+         dataframe["volume_mean_20"] = dataframe["volume"].rolling(20).mean()
+         dataframe["volume_std_20"] = dataframe["volume"].rolling(20).std()
+         dataframe["volume_z_score"] = (dataframe["volume"] - dataframe["volume_mean_20"]) / dataframe["volume_std_20"]
++        # 计算15分钟ATR和3分钟EMA10、EMA20
++        dataframe_15m = self.dp.get_pair_dataframe(pair=pair, timeframe="15m")
++ cuyuan
++        dataframe_15m['atr15m'] = ta.ATR(dataframe_15m, timeperiod=14)
++        dataframe['atr15m'] = dataframe_15m['atr15m'].reindex(dataframe.index, method="ffill")
++        dataframe['ema10'] = ta.EMA(dataframe, timeperiod=10)
++        dataframe['ema20'] = ta.EMA(dataframe, timeperiod=20)
+
+@@ -390,8 +390,11 @@ class FreqaiPrimer(IStrategy):
+         conditions = []
+
+         if "&-price_value_divergence" in dataframe.columns:
+             cond1 = (dataframe["&-price_value_divergence"] > self.sell_threshold * 1.025)
+-            cond2 = (dataframe["rsi"] > 75)
++            rsi_threshold = 70 if trend_score > 50 else 65  # 动态RSI阈值
++            cond2 = (dataframe["rsi"] > rsi_threshold)
++            cond3 = (dataframe['ema10'] < dataframe['ema20'])  # 短期均线交叉
+             sell_condition = cond1 | cond2
+             conditions.append(sell_condition)
+
+@@ -442,6 +442,13 @@ class FreqaiPrimer(IStrategy):
+         initial_stake_amount = trade.stake_amount / 3
+         logger.debug(f"{pair} 首次入场金额: {initial_stake_amount:.2f}, 当前持仓金额: {trade.stake_amount:.2f}, "
+                     f"加仓次数: {trade.nr_of_successful_entries - 1}")
+
++        # 分批出场：短期涨幅触发
++        short_term_return = (current_rate - dataframe['close'].shift(3).iloc[-1]) / dataframe['close'].shift(3).iloc[-1]
++        if short_term_return >= 0.03:  # 3%涨幅
++            reduce_amount = -0.5 * trade.stake_amount  # 卖出50%仓位
++            logger.info(f"{pair} 短期涨幅 {short_term_return*100:.2f}%，卖出50%仓位 {abs(reduce_amount):.2f}")
++            return (reduce_amount, f"Short-term gain {short_term_return*100:.2f}%")
++
+         # 加仓逻辑
+         max_entry_adjustments = self.MAX_ENTRY_POSITION_ADJUSTMENT.value
+@@ -484,12 +484,11 @@ class FreqaiPrimer(IStrategy):
+         # 追踪止损逻辑
+         trailing_stop_start = self.TRAILING_STOP_START.value
+         trailing_stop_distance = self.TRAILING_STOP_DISTANCE.value
+-        # 使用 Sigmoid 映射调整追踪止损参数
+-        sigmoid = 1 / (1 + np.exp(-0.1 * (trend_score - 50)))
+-        trailing_factor = 0.8 + (1.2 - 0.8) * sigmoid  # 牛市(100) -> 1.2, 熊市(0) -> 0.8
+-        distance_factor = 0.7 + (1.5 - 0.7) * sigmoid  # 牛市(100) -> 1.5, 熊市(0) -> 0.7
+-        trailing_stop_start *= trailing_factor
+-        trailing_stop_distance *= distance_factor
++        # 使用15分钟ATR动态调整止损距离
++        atr15m = dataframe['atr15m'].iloc[-1]
++        trailing_stop_distance = min(max(2 * atr15m / current_rate, 0.01), 0.05)  # 2倍ATR，限制在0.01-0.05
++        if trend_score > 50:  # 强趋势时放宽止损
++            trailing_stop_distance *= 1.2
++        trailing_stop_price = max_rate * (1 - trailing_stop_distance)
+
+         if profit_ratio >= trailing_stop_start and not self.trailing_stop_enabled:
+             self.trailing_stop_enabled = True
+             trade.adjust_min_max_rates(current_rate, current_rate)
+             logger.info(f"{pair} 价格上涨超过 {trailing_stop_start*100:.1f}%，启动追踪止损")
+             return None

freqtrade/templates/freqaiprimer.json

@@ -0,0 +1,35 @@
+{
+  "strategy_name": "FreqaiPrimer",
+  "params": {
+    "roi": {},
+    "stoploss": {
+      "stoploss": -0.05
+    },
+    "max_open_trades": {
+      "max_open_trades": 5
+    },
+    "buy": {
+      "ADD_POSITION_THRESHOLD": -0.021,
+      "BUY_THRESHOLD_MAX": -0.001,
+      "BUY_THRESHOLD_MIN": -0.035,
+      "COOLDOWN_PERIOD_MINUTES": 9,
+      "MAX_ENTRY_POSITION_ADJUSTMENT": 3
+    },
+    "sell": {
+      "EXIT_POSITION_RATIO": 0.472,
+      "SELL_THRESHOLD_MAX": 0.065,
+      "SELL_THRESHOLD_MIN": 0.002,
+      "TRAILING_STOP_DISTANCE": 0.015,
+      "TRAILING_STOP_START": 0.016
+    },
+    "protection": {},
+    "trailing": {
+      "trailing_stop": true,
+      "trailing_stop_positive": 0.106,
+      "trailing_stop_positive_offset": 0.196,
+      "trailing_only_offset_is_reached": false
+    }
+  },
+  "ft_stratparam_v": 1,
+  "export_time": "2025-07-01 14:51:29.420394+00:00"
+}

freqtrade/templates/freqaiprimer.json.orig

@@ -0,0 +1,35 @@
+{
+  "strategy_name": "FreqaiPrimer",
+  "params": {
+    "roi": {},
+    "stoploss": {
+      "stoploss": -0.05
+    },
+    "max_open_trades": {
+      "max_open_trades": 5
+    },
+    "buy": {
+      "ADD_POSITION_THRESHOLD": -0.021,
+      "BUY_THRESHOLD_MAX": -0.001,
+      "BUY_THRESHOLD_MIN": -0.035,
+      "COOLDOWN_PERIOD_MINUTES": 9,
+      "MAX_ENTRY_POSITION_ADJUSTMENT": 3
+    },
+    "sell": {
+      "EXIT_POSITION_RATIO": 0.472,
+      "SELL_THRESHOLD_MAX": 0.065,
+      "SELL_THRESHOLD_MIN": 0.002,
+      "TRAILING_STOP_DISTANCE": 0.015,
+      "TRAILING_STOP_START": 0.016
+    },
+    "protection": {},
+    "trailing": {
+      "trailing_stop": true,
+      "trailing_stop_positive": 0.106,
+      "trailing_stop_positive_offset": 0.196,
+      "trailing_only_offset_is_reached": false
+    }
+  },
+  "ft_stratparam_v": 1,
+  "export_time": "2025-07-01 14:51:29.420394+00:00"
+}

freqtrade/templates/freqaiprimer.py

@@ -0,0 +1,830 @@
+import logging
+import numpy as np
+import datetime
+import datetime
+import os
+import json
+import glob
+from functools import reduce
+from freqtrade.persistence import Trade
+import talib.abstract as ta
+from pandas import DataFrame
+from typing import Dict
+from freqtrade.strategy import (DecimalParameter, IStrategy, IntParameter)
+
+logger = logging.getLogger(__name__)
+
+class FreqaiPrimer(IStrategy):
+    """
+    基于 FreqAI 的动态阈值交易策略，集成动态加仓、减仓和自定义 ROI 逻辑，兼容最新 Freqtrade 版本
+    """
+
+    # --- 🧪 Hyperopt Parameters ---
+    TRAILING_STOP_START = DecimalParameter(0.01, 0.05, default=0.03, space='sell', optimize=True)
+    TRAILING_STOP_DISTANCE = DecimalParameter(0.005, 0.02, default=0.01, space='sell', optimize=True)
+
+    BUY_THRESHOLD_MIN = DecimalParameter(-0.15, -0.03, default=-0.07, space='buy', optimize=True)
+    BUY_THRESHOLD_MAX = DecimalParameter(-0.04, -0.005, default=-0.01, space='buy', optimize=True)
+    SELL_THRESHOLD_MIN = DecimalParameter(0.001, 0.02, default=0.005, space='sell', optimize=True)
+    SELL_THRESHOLD_MAX = DecimalParameter(0.02, 0.1, default=0.05, space='sell', optimize=True)
+
+    # 新增：加仓和减仓参数
+    ADD_POSITION_THRESHOLD = DecimalParameter(-0.05, -0.01, default=-0.02, space='buy', optimize=True)
+    EXIT_POSITION_RATIO = DecimalParameter(0.2, 0.7, default=0.5, space='sell', optimize=True)
+    COOLDOWN_PERIOD_MINUTES = IntParameter(1, 10, default=5, space='buy', optimize=True)
+    MAX_ENTRY_POSITION_ADJUSTMENT = IntParameter(1, 3, default=2, space='buy', optimize=True)
+
+    # --- 🛠️ 固定配置参数 ---
+    stoploss = -0.15
+    timeframe = "3m"
+    use_custom_stoploss = True
+    position_adjustment_enable = True  # 启用动态仓位调整
+
+    minimal_roi = {
+        "0": 0.06,    # 30分钟（0-30分钟）内，8% 盈利退出
+        "30": 0.04,   # 2小时（30-120分钟）内，4% 盈利退出
+        "90": 0.025,   # 4小时（120-240分钟）内，2% 盈利退出
+        "270": 0.002    # 8小时（240-480分钟）内，0% 盈利退出
+    }
+
+    plot_config = {
+        "main_plot": {
+            "ema200": {"color": "blue"},
+            "bb_upperband": {"color": "gray"},
+            "bb_lowerband": {"color": "gray"},
+            "bb_middleband": {"color": "gray"}
+        },
+        "subplots": {
+            "Signals": {
+                "enter_long": {"color": "green"},
+                "exit_long": {"color": "red"}
+            },
+            "Price-Value Divergence": {
+                "&-price_value_divergence": {"color": "purple"}
+            },
+            "Volume Z-Score": {
+                "volume_z_score": {"color": "orange"}
+            },
+            "RSI": {
+                "rsi": {"color": "cyan"}
+            }
+        }
+    }
+
+    freqai_info = {
+        "identifier": "test58",
+        "model": "LightGBMRegressor",
+        "feature_parameters": {
+            "include_timeframes": ["3m", "15m", "1h"],
+            "label_period_candles": 12,
+            "include_shifted_candles": 3,
+        },
+        "data_split_parameters": {
+            "test_size": 0.2,
+            "shuffle": False,
+        },
+        "model_training_parameters": {
+            "n_estimators": 200,
+            "learning_rate": 0.05,
+            "num_leaves": 31,
+            "verbose": -1,
+        },
+        "fit_live_predictions_candles": 100,
+        "live_retrain_candles": 100,
+    }
+
+    @staticmethod
+    def linear_map(value, from_min, from_max, to_min, to_max):
+        return (value - from_min) / (from_max - from_min) * (to_max - to_min) + to_min
+
+    def __init__(self, config: dict, *args, **kwargs):
+        super().__init__(config, *args, **kwargs)
+        # 配置日志格式，包含时间戳
+        logging.basicConfig(
+            format='%(asctime)s - %(name)s - %(levelname)s - %(message)s',
+            level=logging.DEBUG,
+            datefmt='%Y-%m-%d %H:%M:%S,%f'[:-3]  # 毫秒精度
+        )
+        logger.debug("✅ 策略已初始化，日志级别设置为 DEBUG")
+        self.trailing_stop_enabled = False
+        self.pair_stats = {}
+        self.stats_logged = False
+        self.fit_live_predictions_candles = self.freqai_info.get("fit_live_predictions_candles", 100)
+        self.last_entry_time = {}  # 记录每个币种的最后入场时间
+
+    def feature_engineering_expand_all(self, dataframe: DataFrame, period: int, metadata: dict, **kwargs) -> DataFrame:
+        dataframe["%-rsi-period"] = ta.RSI(dataframe, timeperiod=period)
+        dataframe["%-sma-period"] = ta.SMA(dataframe, timeperiod=period)
+        dataframe["%-ema-period"] = ta.EMA(dataframe, timeperiod=period)
+
+        real = ta.TYPPRICE(dataframe)
+        upperband, middleband, lowerband = ta.BBANDS(real, timeperiod=period, nbdevup=2.0, nbdevdn=2.0)
+        dataframe["bb_lowerband-period"] = lowerband
+        dataframe["bb_upperband-period"] = upperband
+        dataframe["bb_middleband-period"] = middleband
+        dataframe["%-bb_width-period"] = (dataframe["bb_upperband-period"] - dataframe["bb_lowerband-period"]) / dataframe["bb_middleband-period"]
+
+        dataframe["%-mfi-period"] = ta.MFI(dataframe, timeperiod=period)
+        dataframe["%-adx-period"] = ta.ADX(dataframe, timeperiod=period)
+        dataframe["%-relative_volume-period"] = dataframe["volume"] / dataframe["volume"].rolling(period).mean()
+
+        dataframe["ema200"] = ta.EMA(dataframe, timeperiod=200)
+        dataframe["%-price_value_divergence"] = (dataframe["close"] - dataframe["ema200"]) / dataframe["ema200"]
+
+        columns_to_clean = [
+            "%-rsi-period", "%-mfi-period", "%-sma-period", "%-ema-period", "%-adx-period",
+            "bb_lowerband-period", "bb_middleband-period", "bb_upperband-period",
+            "%-bb_width-period", "%-relative_volume-period", "%-price_value_divergence"
+        ]
+        for col in columns_to_clean:
+            dataframe[col] = dataframe[col].replace([np.inf, -np.inf], 0).ffill().fillna(0)
+
+        pair = metadata.get('pair', 'Unknown')
+        logger.debug(f"[{pair}] 特征工程完成，列：{list(dataframe.columns)}")
+        return dataframe
+
+    def set_freqai_targets(self, dataframe: DataFrame, metadata: dict, **kwargs) -> DataFrame:
+        pair = metadata.get('pair', 'Unknown')
+        if len(dataframe) < 200:
+            logger.warning(f"[{pair}] 数据量不足（{len(dataframe)}根K线），需要至少200根K线进行训练")
+            return dataframe
+
+        dataframe["ema200"] = ta.EMA(dataframe, timeperiod=200)
+        dataframe["&-price_value_divergence"] = (dataframe["close"] - dataframe["ema200"]) / dataframe["ema200"]
+
+        dataframe["volume_mean_20"] = dataframe["volume"].rolling(20).mean()
+        dataframe["volume_std_20"] = dataframe["volume"].rolling(20).std()
+        dataframe["volume_z_score"] = (dataframe["volume"] - dataframe["volume_mean_20"]) / dataframe["volume_std_20"]
+
+        dataframe["&-price_value_divergence"] = dataframe["&-price_value_divergence"].replace([np.inf, -np.inf], 0).ffill().fillna(0)
+        dataframe["volume_z_score"] = dataframe["volume_z_score"].replace([np.inf, -np.inf], 0).ffill().fillna(0)
+
+        return dataframe
+
+    def populate_indicators(self, dataframe: DataFrame, metadata: dict) -> DataFrame:
+        pair = metadata.get('pair', 'Unknown')
+        logger.info(f"[{pair}] 当前可用列（调用FreqAI前）：{list(dataframe.columns)}")
+
+        # 计算200周期EMA和历史价值背离
+        dataframe["ema200"] = ta.EMA(dataframe, timeperiod=200)
+        dataframe["price_value_divergence"] = (dataframe["close"] - dataframe["ema200"]) / dataframe["ema200"]
+
+        # 调用FreqAI预测价值背离
+        if not hasattr(self, 'freqai') or self.freqai is None:
+            logger.error(f"[{pair}] FreqAI 未初始化，请确保回测命令中启用了 --freqai")
+            dataframe["&-price_value_divergence"] = dataframe["price_value_divergence"]
+        else:
+            dataframe = self.freqai.start(dataframe, metadata, self)
+            if "&-price_value_divergence" not in dataframe.columns:
+                logger.warning(f"[{pair}] 回归模型未生成 &-price_value_divergence，回退到规则计算")
+                dataframe["&-price_value_divergence"] = dataframe["price_value_divergence"]
+
+        # 计算 custom_exit 和其他方法所需的指标
+        dataframe["rsi"] = ta.RSI(dataframe, timeperiod=14)
+        stochrsi = ta.STOCHRSI(dataframe, timeperiod=14, fastk_period=3, fastd_period=3)
+        dataframe["stochrsi_k"] = stochrsi["fastk"]
+        dataframe["adx"] = ta.ADX(dataframe, timeperiod=14)
+        dataframe["short_term_return"] = dataframe["close"].pct_change(5) * 100  # 5根K线的百分比回报
+        dataframe["volume_mean_20"] = dataframe["volume"].rolling(20).mean()
+        dataframe["volume_std_20"] = dataframe["volume"].rolling(20).std()
+        dataframe["volume_z_score"] = (dataframe["volume"] - dataframe["volume_mean_20"]) / dataframe["volume_std_20"].replace(0, np.nan)
+
+        # 计算布林带（用于其他逻辑）
+        upperband, middleband, lowerband = ta.BBANDS(dataframe["close"], timeperiod=20, nbdevup=2.0, nbdevdn=2.0)
+        dataframe["bb_upperband"] = upperband
+        dataframe["bb_middleband"] = middleband
+        dataframe["bb_lowerband"] = lowerband
+
+        # 数据清理
+        for col in ["ema200", "bb_upperband", "bb_middleband", "bb_lowerband", "rsi", "stochrsi_k", 
+                    "adx", "short_term_return", "volume_z_score", "&-price_value_divergence", 
+                    "price_value_divergence"]:
+            dataframe[col] = dataframe[col].replace([np.inf, -np.inf], 0).ffill().fillna(0)
+
+        # 调试日志
+        logger.debug(f"[{pair}] 最新数据 - close: {dataframe['close'].iloc[-1]:.6f}, "
+                    f"rsi: {dataframe['rsi'].iloc[-1]:.2f}, "
+                    f"stochrsi_k: {dataframe['stochrsi_k'].iloc[-1]:.2f}, "
+                    f"adx: {dataframe['adx'].iloc[-1]:.2f}, "
+                    f"short_term_return: {dataframe['short_term_return'].iloc[-1]:.2f}%, "
+                    f"volume_z_score: {dataframe['volume_z_score'].iloc[-1]:.2f}, "
+                    f"&-price_value_divergence: {dataframe['&-price_value_divergence'].iloc[-1]:.6f}")
+
+        # 获取 labels_mean 和 labels_std
+        labels_mean = None
+        labels_std = None
+        try:
+            model_base_dir = os.path.join(self.config["user_data_dir"], "models", self.freqai_info["identifier"])
+            pair_base = pair.split('/')[0] if '/' in pair else pair
+            sub_dirs = glob.glob(os.path.join(model_base_dir, f"sub-train-{pair_base}_*"))
+            if sub_dirs:
+                latest_sub_dir = max(sub_dirs, key=lambda x: int(x.split('_')[-1]))
+                pair_base_lower = pair_base.lower()
+                timestamp = latest_sub_dir.split('_')[-1]
+                metadata_file = os.path.join(latest_sub_dir, f"cb_{pair_base_lower}_{timestamp}_metadata.json")
+                if os.path.exists(metadata_file):
+                    with open(metadata_file, "r") as f:
+                        metadata = json.load(f)
+                        labels_mean = metadata["labels_mean"]["&-price_value_divergence"]
+                        labels_std = metadata["labels_std"]["&-price_value_divergence"]
+                        logger.info(f"[{pair}] 从最新子目录 {latest_sub_dir} 读取 labels_mean: {labels_mean}, labels_std: {labels_std}")
+                else:
+                    logger.warning(f"[{pair}] 最新的 metadata.json 文件 {metadata_file} 不存在")
+        except Exception as e:
+            logger.warning(f"[{pair}] 无法从子目录读取 labels_mean 和 labels_std: {e}")
+
+        if labels_mean is None or labels_std is None:
+            logger.warning(f"[{pair}] 无法获取 labels_mean 和 labels_std，重新计算")
+            dataframe["&-price_value_divergence_actual"] = (dataframe["close"] - dataframe["ema200"]) / dataframe["ema200"]
+            dataframe["&-price_value_divergence_actual"] = dataframe["&-price_value_divergence_actual"].replace([np.inf, -np.inf], 0).ffill().fillna(0)
+            recent_data = dataframe["&-price_value_divergence_actual"].tail(self.fit_live_predictions_candles)
+            labels_mean = recent_data.mean()
+            labels_std = recent_data.std() or 0.01  # 防止标准差为 0
+            logger.warning(f"[{pair}] labels_std 计算异常，使用默认值 0.01")
+
+        # 动态调整买卖阈值
+        market_trend_score = self.get_market_trend(dataframe=dataframe, metadata={'pair': pair})
+        k_buy = self.linear_map(market_trend_score, 0, 100, 1.2, 0.8)
+        k_sell = self.linear_map(market_trend_score, 0, 100, 1.5, 1.0)
+        self.buy_threshold = max(min(labels_mean - k_buy * labels_std, self.BUY_THRESHOLD_MAX.value), self.BUY_THRESHOLD_MIN.value)
+        self.sell_threshold = max(min(labels_mean + k_sell * labels_std, self.SELL_THRESHOLD_MAX.value), self.SELL_THRESHOLD_MIN.value)
+
+        logger.info(f"[{pair}] 市场趋势得分: {market_trend_score:.2f}, labels_mean: {labels_mean:.4f}, labels_std: {labels_std:.4f}, "
+                    f"buy_threshold: {self.buy_threshold:.4f}, sell_threshold: {self.sell_threshold:.4f}")
+
+        return dataframe
+
+    def trailing_space(self):
+        return [
+            DecimalParameter(0.01, 0.05, name="trailing_stop_start"),
+            DecimalParameter(0.005, 0.02, name="trailing_stop_distance")
+        ]
+
+    def leverage_space(self):
+        return [
+            DecimalParameter(-0.05, -0.01, name="add_position_threshold", default=-0.02),
+            DecimalParameter(0.2, 0.7, name="exit_position_ratio", default=0.5),
+            IntParameter(1, 10, name="cooldown_period_minutes", default=5),
+            IntParameter(1, 3, name="max_entry_position_adjustment", default=2)
+        ]
+
+    def populate_entry_trend(self, dataframe: DataFrame, metadata: dict) -> DataFrame:
+        pair = metadata.get('pair', 'Unknown')
+        conditions = []
+
+        # 获取市场趋势得分
+        trend_score = self.get_market_trend(dataframe=dataframe, metadata=metadata)
+
+        # 动态调整成交量阈值：牛市（trend_score=100）-> 0.5，熊市（trend_score=0）-> 1.5
+        volume_z_score_min = 0.5
+        volume_z_score_max = 1.5
+        volume_z_score_threshold = self.linear_map(trend_score, 0, 100, volume_z_score_max, volume_z_score_min)
+
+        # 动态调整 RSI 阈值：牛市（trend_score=100）-> 40，熊市（trend_score=0）-> 60
+        rsi_min = 35
+        rsi_max = 55
+        rsi_threshold = self.linear_map(trend_score, 0, 100, rsi_max, rsi_min)
+
+        # 新增：动态调整 STOCHRSI 阈值，牛市 -> 30，熊市 -> 50
+        stochrsi_min = 25
+        stochrsi_max = 45
+        stochrsi_threshold = self.linear_map(trend_score, 0, 100, stochrsi_max, stochrsi_min)
+
+        if "&-price_value_divergence" in dataframe.columns:
+            # 计算 STOCHRSI
+            stochrsi = ta.STOCHRSI(dataframe, timeperiod=14, fastk_period=3, fastd_period=3)
+            dataframe["stochrsi_k"] = stochrsi["fastk"]
+            # 添加 15m 时间框架的 EMA200
+            dataframe_15m = self.dp.get_pair_dataframe(pair=pair, timeframe="15m")
+            dataframe_15m["ema200_15m"] = ta.EMA(dataframe_15m, timeperiod=200)
+            dataframe["ema200_15m"] = dataframe_15m["ema200_15m"].reindex(dataframe.index, method="ffill")
+
+            cond1 = (dataframe["&-price_value_divergence"] < self.buy_threshold)
+            cond2 = (dataframe["volume_z_score"] > volume_z_score_threshold)
+            cond3 = (dataframe["rsi"] < rsi_threshold)
+            cond4 = (dataframe["close"] <= dataframe["bb_lowerband"])
+            cond5 = (dataframe["stochrsi_k"] < stochrsi_threshold)
+            buy_condition = cond1 & cond2 & cond3 & cond4 & cond5
+
+            conditions.append(buy_condition)
+
+            divergence_value = dataframe['&-price_value_divergence'].iloc[-1] if not dataframe['&-price_value_divergence'].isna().all() else np.nan
+            volume_z_score_value = dataframe['volume_z_score'].iloc[-1] if not dataframe['volume_z_score'].isna().all() else np.nan
+            rsi_value = dataframe['rsi'].iloc[-1] if not dataframe['rsi'].isna().all() else np.nan
+            stochrsi_value = dataframe['stochrsi_k'].iloc[-1] if not dataframe['stochrsi_k'].isna().all() else np.nan
+            logger.debug(f"[{pair}] 买入条件检查 - "
+                        f"&-price_value_divergence={divergence_value:.6f} < {self.buy_threshold:.6f}: {cond1.iloc[-1]}, "
+                        f"volume_z_score={volume_z_score_value:.2f} > {volume_z_score_threshold:.2f}: {cond2.iloc[-1]}, "
+                        f"rsi={rsi_value:.2f} < {rsi_threshold:.2f}: {cond3.iloc[-1]}, "
+                        f"close={dataframe['close'].iloc[-1]:.6f} <= bb_lowerband={dataframe['bb_lowerband'].iloc[-1]:.6f}: {cond4.iloc[-1]}, "
+                        f"stochrsi_k={stochrsi_value:.2f} < {stochrsi_threshold:.2f}: {cond5.iloc[-1]}")
+        else:
+            logger.warning(f"[{pair}] ⚠️ &-price_value_divergence 列缺失，跳过买入信号生成")
+
+        if len(conditions) > 0:
+            combined_condition = reduce(lambda x, y: x & y, conditions)
+            if combined_condition.any():
+                dataframe.loc[combined_condition, 'enter_long'] = 1
+                logger.info(f"[{pair}] 买入信号触发，条件满足，趋势得分：{trend_score:.2f}")
+            else:
+                logger.debug(f"[{pair}] 买入条件未满足，无买入信号")
+        else:
+            logger.debug(f"[{pair}] 无有效买入条件")
+
+        return dataframe
+
+    def populate_exit_trend(self, dataframe: DataFrame, metadata: dict) -> DataFrame:
+        pair = metadata.get('pair', 'Unknown')
+        conditions = []
+
+        if "&-price_value_divergence" in dataframe.columns:
+            # 计算额外指标：ADX（仅用于 cond1）
+            dataframe["adx"] = ta.ADX(dataframe, timeperiod=14)
+
+            # 获取市场趋势得分
+            trend_score = self.get_market_trend(dataframe=dataframe, metadata={'pair': pair})
+
+            # 条件 1：高阈值 &-price_value_divergence
+            cond1 = (
+                (dataframe["&-price_value_divergence"] > self.sell_threshold * 1.06) &
+                (dataframe["adx"] > 20)
+            )
+            conditions.append(cond1)
+
+            # 调试日志
+            divergence_value = dataframe["&-price_value_divergence"].iloc[-1] if not dataframe["&-price_value_divergence"].isna().all() else np.nan
+            adx_value = dataframe["adx"].iloc[-1] if not dataframe["adx"].isna().all() else np.nan
+            logger.debug(f"[{pair}] 卖出条件检查 - "
+                        f"&-price_value_divergence={divergence_value:.6f} > {self.sell_threshold * 1.06:.6f}: {cond1.iloc[-1]}, "
+                        f"adx={adx_value:.2f}, trend_score={trend_score:.2f}")
+        else:
+            logger.warning(f"[{pair}] ⚠️ &-price_value_divergence 列缺失，跳过该条件")
+
+        if len(conditions) > 0:
+            dataframe.loc[reduce(lambda x, y: x & y, conditions), 'exit_long'] = 1
+            logger.info(f"[{pair}] 出场信号触发（高背离），条件满足，趋势得分：{trend_score:.2f}")
+        else:
+            logger.debug(f"[{pair}] 无有效卖出条件")
+
+        return dataframe
+
+
+    def buy_space(self):
+        return [
+            DecimalParameter(-0.1, -0.01, name="buy_threshold_min"),
+            DecimalParameter(-0.02, -0.001, name="buy_threshold_max"),
+            DecimalParameter(-0.05, -0.01, name="add_position_threshold", default=-0.02),
+            IntParameter(1, 10, name="cooldown_period_minutes", default=5),
+            IntParameter(1, 3, name="max_entry_position_adjustment", default=2)
+        ]
+
+    def sell_space(self):
+        return [
+            DecimalParameter(0.001, 0.02, name="sell_threshold_min"),
+            DecimalParameter(0.02, 0.1, name="sell_threshold_max"),
+            DecimalParameter(0.2, 0.7, name="exit_position_ratio", default=0.5)
+        ]
+
+    def adjust_trade_position(self, trade: Trade, current_time: datetime,
+                            current_rate: float, current_profit: float,
+                            min_stake: float | None, max_stake: float,
+                            current_entry_rate: float, current_exit_rate: float,
+                            current_entry_profit: float, current_exit_profit: float,
+                            **kwargs) -> float | None | tuple[float | None, str | None]:
+        """
+        动态调整仓位：支持加仓、减仓、追踪止损和最大持仓时间限制
+        参数：
+        - trade: 当前交易对象
+        - current_time: 当前时间
+        - current_rate: 当前价格
+        - current_profit: 当前总盈利
+        - min_stake: 最小下注金额
+        - max_stake: 最大下注金额
+        - current_entry_rate: 当前入场价格
+        - current_exit_rate: 当前退出价格
+        - current_entry_profit: 当前入场盈利
+        - current_exit_profit: 当前退出盈利
+        返回：
+        - 调整金额（正数为加仓，负数为减仓）或 None
+        """
+        pair = trade.pair
+        dataframe = self.dp.get_pair_dataframe(pair, self.timeframe)
+        trend_score = self.get_market_trend(dataframe=dataframe, metadata={'pair': pair})
+        hold_time = (current_time - trade.open_date_utc).total_seconds() / 60
+        profit_ratio = (current_rate - trade.open_rate) / trade.open_rate
+
+        initial_stake_amount = trade.stake_amount / 3 * 2
+        logger.debug(f"{pair} 首次入场金额: {initial_stake_amount:.2f}, 当前持仓金额: {trade.stake_amount:.2f}, "
+                    f"加仓次数: {trade.nr_of_successful_entries - 1}")
+
+        # 加仓逻辑
+        max_entry_adjustments = self.MAX_ENTRY_POSITION_ADJUSTMENT.value
+        if trade.nr_of_successful_entries <= max_entry_adjustments + 1:
+            add_position_threshold = self.ADD_POSITION_THRESHOLD.value
+            # 线性映射加仓阈值，趋势值越高，加仓越严格
+            add_threshold = 80 - 30 * (trend_score / 100)  # 趋势值 100 -> 50, 0 -> 80
+            if profit_ratio <= add_position_threshold and hold_time > 5 and trend_score <= add_threshold:
+                logger.debug(f"{pair} 初始下注金额: {initial_stake_amount:.2f}, trend_score: {trend_score:.2f}, add_threshold: {add_threshold} ")
+
+                # 计算加仓金额
+                add_count = trade.nr_of_successful_entries - 1
+                multipliers = [2, 4, 8, 16, 32]
+                if add_count < len(multipliers):
+                    multiplier = multipliers[add_count]
+                    add_amount = initial_stake_amount * multiplier
+                    logger.debug(f"{pair} 第 {add_count + 1} 次加仓，倍数={multiplier}, "
+                                f"金额 = {initial_stake_amount:.2f} * {multiplier} = {add_amount:.2f}")
+                    logger.debug(f"{pair} 加仓计算: 第 {add_count + 1} 次加仓，倍数={multiplier}, "
+                                f"金额 = {initial_stake_amount:.2f} * {multiplier} = {add_amount:.2f}")
+
+                    if min_stake is not None and add_amount < min_stake:
+                        logger.warning(f"{pair} 加仓金额 {add_amount:.2f} 低于最小下注金额 {min_stake:.2f}，取消加仓")
+                        return (None, f"Add amount {add_amount:.2f} below min_stake {min_stake:.2f}")
+                    if add_amount > max_stake:
+                        logger.warning(f"{pair} 加仓金额 {add_amount:.2f} 超出最大可用金额 {max_stake:.2f}，调整为 {max_stake:.2f}")
+                        add_amount = max_stake
+                    logger.info(f"{pair} 价格下跌 {profit_ratio*100:.2f}%，触发第 {add_count + 1} 次加仓 {add_amount:.2f}")
+                    return (add_amount, f"Price dropped {profit_ratio*100:.2f}%, add {add_amount:.2f}")
+
+        # 减仓逻辑
+        exit_position_ratio = self.EXIT_POSITION_RATIO.value
+        if profit_ratio >= 0.03:
+            # 趋势值越高，减仓比例越低
+            reduce_factor = 0.6 + 0.4 * (1 - trend_score / 100)  # 牛市(100) -> 0.6, 熊市(0) -> 1.0
+            reduce_amount = -exit_position_ratio * reduce_factor * trade.stake_amount
+            logger.info(f"{pair} 趋势值 {trend_score:.2f}，利润 {profit_ratio*100:.2f}%，减仓 {abs(reduce_amount):.2f}")
+            return (reduce_amount, f"Profit {profit_ratio*100:.2f}%")
+        elif profit_ratio >= 0.05:
+            reduce_factor = 1.4 - 0.4 * (trend_score / 100)  # 牛市(100) -> 1.0, 熊市(0) -> 1.4
+            reduce_amount = -exit_position_ratio * reduce_factor * trade.stake_amount
+            logger.info(f"{pair} 趋势值 {trend_score:.2f}，利润 {profit_ratio*100:.2f}%，减仓 {abs(reduce_amount):.2f}")
+            return (reduce_amount, f"Profit {profit_ratio*100:.2f}%")
+
+        # 追踪止损逻辑
+        trailing_stop_start = self.TRAILING_STOP_START.value
+        trailing_stop_distance = self.TRAILING_STOP_DISTANCE.value
+        # 使用 Sigmoid 映射调整追踪止损参数
+        sigmoid = 1 / (1 + np.exp(-0.1 * (trend_score - 50)))
+        trailing_factor = 0.8 + (1.2 - 0.8) * sigmoid  # 牛市(100) -> 1.2, 熊市(0) -> 0.8
+        distance_factor = 0.7 + (1.5 - 0.7) * sigmoid  # 牛市(100) -> 1.5, 熊市(0) -> 0.7
+        trailing_stop_start *= trailing_factor
+        trailing_stop_distance *= distance_factor
+
+        if profit_ratio >= trailing_stop_start and not self.trailing_stop_enabled:
+            self.trailing_stop_enabled = True
+            trade.adjust_min_max_rates(current_rate, current_rate)
+            logger.info(f"{pair} 价格上涨超过 {trailing_stop_start*100:.1f}%，启动追踪止损")
+            return None
+
+        if self.trailing_stop_enabled:
+            max_rate = trade.max_rate or current_rate
+            trailing_stop_price = max_rate * (1 - trailing_stop_distance)
+            if current_rate < trailing_stop_price:
+                logger.info(f"{pair} 价格回落至 {trailing_stop_price:.6f}，触发全部卖出")
+                return (-trade.stake_amount, f"Trailing stop at {trailing_stop_price:.6f}")
+            trade.adjust_min_max_rates(current_rate, trade.min_rate)
+            return None
+
+        return None
+
+    def confirm_trade_entry(self, pair: str, order_type: str, amount: float, rate: float,
+                            time_in_force: str, current_time: datetime, **kwargs) -> bool:
+        # 调试日志：记录输入参数
+        logger.debug(f"[{pair}] confirm_trade_entry called with rate={rate}, type(rate)={type(rate)}, "
+                    f"amount={amount}, order_type={order_type}, time_in_force={time_in_force}")
+
+        # 检查 rate 是否有效
+        if not isinstance(rate, (float, int)) or rate is None:
+            logger.error(f"[{pair}] Invalid rate value: {rate} (type: {type(rate)}). Skipping trade entry.")
+            return False
+
+        market_trend_score = self.get_market_trend(dataframe=DataFrame, metadata={'pair': pair})
+        cooldown_period_minutes = self.COOLDOWN_PERIOD_MINUTES.value if market_trend_score > 50 else self.COOLDOWN_PERIOD_MINUTES.value // 2
+
+        if pair in self.last_entry_time:
+            last_time = self.last_entry_time[pair]
+            if (current_time - last_time).total_seconds() < cooldown_period_minutes * 60:
+                logger.info(f"[{pair}] 冷却期内（{cooldown_period_minutes} 分钟），跳过本次入场")
+                return False
+
+        self.last_entry_time[pair] = current_time
+        self.trailing_stop_enabled = False
+        try:
+            logger.info(f"[{pair}] 确认入场，价格：{float(rate):.6f}")
+        except (ValueError, TypeError) as e:
+            logger.error(f"[{pair}] Failed to format rate: {rate} (type: {type(rate)}), error: {e}")
+            return False
+
+        return True
+
+    def confirm_trade_exit(self, pair: str, trade: Trade, order_type: str, amount: float,
+             rate: float, time_in_force: str, exit_reason: str,
+             current_time: datetime, **kwargs) -> bool:
+        adjusted_rate = rate * (1 + 0.0025)
+        logger.info(f"[{pair}] 退出交易，原因：{exit_reason}, 原始利润：{trade.calc_profit_ratio(rate):.2%},"f"调整后卖出价：{adjusted_rate:.6f}")
+        return True
+
+    def custom_roi(self, trade: Trade, current_profit: float, current_time: datetime, trade_dur: int, 
+                   current_rate: float = None, min_stake: float | None = None, max_stake: float | None = None) -> dict:
+        """
+        动态调整 ROI 表格，基于 FreqAI 预测的 &-price_value_divergence 和 RSI。
+        - 负的 divergence（预测上涨）或低 RSI 时提高 ROI。
+        - 正的 divergence（预测下跌）或高 RSI 时降低 ROI。
+        - 长时间持仓降低 ROI 目标。
+        """
+        pair = trade.pair
+        logger.debug(f"[{pair}] 计算自定义 ROI，当前盈利: {current_profit:.2%}, 持仓时间: {trade_dur} 分钟")
+
+        # 获取最新数据
+        dataframe = self.dp.get_pair_dataframe(pair=pair, timeframe=self.timeframe)
+        dataframe = self.populate_indicators(dataframe, {'pair': pair})  # 计算指标
+
+        # 获取 FreqAI 预测和 RSI
+        divergence = dataframe["&-price_value_divergence"].iloc[-1] if "&-price_value_divergence" in dataframe else 0
+        rsi = dataframe["rsi"].iloc[-1] if "rsi" in dataframe else 50
+
+        # 计算调整系数
+        # 1. Divergence 调整：负值（预测上涨）-> 提高 ROI，正值（预测下跌）-> 降低 ROI
+        divergence_factor = self.linear_map(divergence, -0.1, 0.1, 1.2, 0.8)
+        
+        # 2. RSI 调整：低 RSI（超卖）-> 提高 ROI，高 RSI（超买）-> 降低 ROI
+        rsi_factor = self.linear_map(rsi, 30, 70, 1.2, 0.8)
+        
+        # 3. 时间调整：持仓时间越长，ROI 目标降低
+        time_factor = self.linear_map(trade_dur, 0, 240, 1.0, 0.7)  # 4小时后 ROI 降低到 70%
+
+        # 综合调整系数
+        roi_factor = divergence_factor * rsi_factor * time_factor
+
+        # 默认 ROI 表格
+        base_roi = {
+            0: 0.06,
+            30: 0.04,
+            90: 0.025,
+            270: 0.002
+        }
+
+        # 动态调整 ROI，限制在 0% 到 20% 之间
+        dynamic_roi = {time: min(max(roi * roi_factor, 0.0), 0.2) for time, roi in base_roi.items()}
+        
+        logger.debug(f"[{pair}] Divergence: {divergence:.4f}, RSI: {rsi:.2f}, 持仓时间: {trade_dur} 分钟, "
+                     f"调整系数: divergence={divergence_factor:.2f}, rsi={rsi_factor:.2f}, time={time_factor:.2f}, "
+                     f"总系数={roi_factor:.2f}, 动态 ROI 表格: {dynamic_roi}")
+        return dynamic_roi
+
+    def custom_entry_price(self, pair: str, trade: Trade | None, current_time: datetime, proposed_rate: float,
+                           entry_tag: str | None, side: str, **kwargs) -> float:
+        adjusted_rate = proposed_rate * (1 - 0.005)
+        logger.debug(f"[{pair}] 自定义买入价：{adjusted_rate:.6f}（原价：{proposed_rate:.6f}）")
+        return adjusted_rate
+
+    def custom_exit(self, pair: str, trade: Trade, current_time: datetime, current_rate: float,
+                    current_profit: float, **kwargs) -> tuple[bool, str] | None:
+        """
+        自定义退出逻辑，检查超买信号和快速拉升信号，确保盈利不低于动态 ROI 预期。
+        避免调用 populate_indicators，直接使用预计算的指标。
+        返回：(bool, str) 表示是否退出及退出原因，或 None 表示不退出。
+        """
+        logger.debug(f"[{pair}] 检查自定义退出，当前盈利: {current_profit:.2%}")
+
+        # 获取最新数据
+        dataframe = self.dp.get_pair_dataframe(pair=pair, timeframe=self.timeframe)
+
+        # 验证所需指标
+        required_columns = ["rsi", "stochrsi_k", "adx", "short_term_return", "volume_z_score", "close"]
+        missing_columns = [col for col in required_columns if col not in dataframe.columns]
+        if missing_columns:
+            logger.warning(f"[{pair}] 缺少必要指标: {missing_columns}，跳过自定义退出")
+            return None
+
+        # 计算持仓时间（分钟）
+        trade_dur = (current_time - trade.open_date).total_seconds() / 60
+
+        # 获取动态 ROI 表格
+        dynamic_roi = self.custom_roi(
+            trade=trade,
+            current_profit=current_profit,
+            current_time=current_time,
+            trade_dur=trade_dur,
+            current_rate=current_rate
+        )
+
+        # 查找当前持仓时间的预期 ROI
+        expected_roi = 0.0
+        for time, roi in sorted(dynamic_roi.items()):
+            if trade_dur >= time:
+                expected_roi = roi
+            else:
+                break
+        logger.debug(f"[{pair}] 当前持仓时间: {trade_dur:.2f} 分钟, 预期 ROI: {expected_roi:.4f}, 当前盈利: {current_profit:.4f}")
+
+        # 获取市场趋势得分
+        trend_score = self.get_market_trend(dataframe=dataframe, metadata={'pair': pair})
+
+        # 定义退出条件
+        conditions = []
+        exit_reasons = []
+
+        # 使用 for 循环检查条件
+        exit_checks = [
+            {  # 条件 2：超买信号
+                "name": "超买信号",
+                "condition": (
+                    (dataframe["rsi"].iloc[-1] > 65) &
+                    (dataframe["stochrsi_k"].iloc[-1] > 70) &
+                    (dataframe["adx"].iloc[-1] > 25) &
+                    (current_profit >= expected_roi)
+                ),
+                "details": f"RSI={dataframe['rsi'].iloc[-1]:.2f}, StochRSI={dataframe['stochrsi_k'].iloc[-1]:.2f}, "
+                        f"ADX={dataframe['adx'].iloc[-1]:.2f}, 盈利={current_profit:.4f}>={expected_roi:.4f}"
+            },
+            {  # 条件 3：快速拉升
+                "name": "快速拉升",
+                "condition": (
+                    (dataframe["short_term_return"].iloc[-1] > self.linear_map(trend_score, 0, 100, 3.5, 2.0)) &
+                    (dataframe["close"].iloc[-1] / dataframe["close"].shift(5).iloc[-1] - 1 > 0.02) &
+                    (dataframe["stochrsi_k"].iloc[-1] > 70) &
+                    (dataframe["volume_z_score"].iloc[-1] > 1.0) &
+                    (current_profit >= expected_roi)
+                ),
+                "details": f"短期涨幅={dataframe['short_term_return'].iloc[-1]:.2f}%, "
+                        f"盈利={current_profit:.4f}>={expected_roi:.4f}, "
+                        f"VolumeZ={dataframe['volume_z_score'].iloc[-1]:.2f}"
+            }
+        ]
+
+        for check in exit_checks:
+            if check["condition"]:
+                conditions.append(True)
+                exit_reasons.append(f"{check['name']} ({check['details']})")
+
+        # 调试日志
+        rsi_value = dataframe["rsi"].iloc[-1] if not dataframe["rsi"].isna().all() else np.nan
+        stochrsi_value = dataframe["stochrsi_k"].iloc[-1] if not dataframe["stochrsi_k"].isna().all() else np.nan
+        adx_value = dataframe["adx"].iloc[-1] if not dataframe["adx"].isna().all() else np.nan
+        short_term_return = dataframe["short_term_return"].iloc[-1] if not dataframe["short_term_return"].isna().all() else np.nan
+        volume_z_score = dataframe["volume_z_score"].iloc[-1] if not dataframe["volume_z_score"].isna().all() else np.nan
+        rapid_rise_threshold = self.linear_map(trend_score, 0, 100, 3.5, 2.0)
+        logger.debug(f"[{pair}] 自定义退出检查 - "
+                    f"rsi={rsi_value:.2f}>65 & stochrsi_k={stochrsi_value:.2f}>70 & adx={adx_value:.2f}>25 & profit>={expected_roi:.4f}: {exit_checks[0]['condition']}, "
+                    f"short_term_return={short_term_return:.2f}%>{rapid_rise_threshold:.2f}% & profit>=2.00% & volume_z_score={volume_z_score:.2f}>1.0 & profit>={expected_roi:.4f}: {exit_checks[1]['condition']}, "
+                    f"trend_score={trend_score:.2f}, current_profit={current_profit:.4f}")
+
+        # 组合逻辑
+        if trend_score > 85:
+            # 牛市：需要高背离 (exit_long == 1) 和至少一个条件，或两个条件
+            if len(conditions) >= 1 and dataframe["exit_long"].iloc[-1] == 1:
+                reason = "牛市退出: 高背离 + " + " & ".join(exit_reasons)
+                logger.info(f"[{pair}] 触发自定义退出，原因: {reason}")
+                return True, reason
+            elif len(conditions) >= 2:
+                reason = "牛市退出: " + " & ".join(exit_reasons)
+                logger.info(f"[{pair}] 触发自定义退出，原因: {reason}")
+                return True, reason
+        else:
+            # 中性或熊市：任一条件满足
+            if len(conditions) >= 1:
+                reason = "退出: " + " & ".join(exit_reasons)
+                logger.info(f"[{pair}] 触发自定义退出，原因: {reason}")
+                return True, reason
+
+        logger.debug(f"[{pair}] 无自定义退出条件满足")
+        return None
+
+    def custom_exit_price(self, pair: str, trade: Trade,
+                          current_time: datetime, proposed_rate: float,
+                          current_profit: float, exit_tag: str | None, **kwargs) -> float:
+        adjusted_rate = proposed_rate * (1 + 0.0025)
+        logger.debug(f"[{pair}] 自定义卖出价：{adjusted_rate:.6f}（原价：{proposed_rate:.6f}）")
+        return adjusted_rate
+
+    def get_market_trend(self, dataframe: DataFrame = None, metadata: dict = None) -> int:
+        try:
+            timeframes = ["3m", "15m", "1h"]
+            weights = {"3m": 0.3, "15m": 0.35, "1h": 0.35}
+            trend_scores = {}
+            pair = metadata.get('pair', 'Unknown') if metadata else 'Unknown'
+
+            # 检查 pair 是否有效
+            if pair == 'Unknown':
+                logger.error(f"[{pair}] Invalid pair in metadata: {metadata}. Returning default score 50")
+                return 50
+
+            logger.debug(f"[{pair}] 正在计算多时间框架市场趋势得分")
+
+            for tf in timeframes:
+                # 优先使用传入的 dataframe（如果匹配主时间框架），否则加载目标币对数据
+                pair_df = dataframe if tf == self.timeframe and dataframe is not None else self.dp.get_pair_dataframe(pair=pair, timeframe=tf)
+
+                min_candles = 200 if tf == "3m" else 100 if tf == "15m" else 50
+                if pair_df.empty or len(pair_df) < min_candles:
+                    logger.warning(f"[{pair}] 数据不足（{tf}），使用默认得分：50")
+                    trend_scores[tf] = 50
+                    continue
+
+                # 价格趋势
+                ema_short_period = 50 if tf == "3m" else 20 if tf == "15m" else 12
+                ema_long_period = 200 if tf == "3m" else 80 if tf == "15m" else 50
+                pair_df["ema_short"] = ta.EMA(pair_df, timeperiod=ema_short_period)
+                pair_df["ema_long"] = ta.EMA(pair_df, timeperiod=ema_long_period)
+                pair_df["ema_short_slope"] = (pair_df["ema_short"] - pair_df["ema_short"].shift(10)) / pair_df["ema_short"].shift(10)
+
+                price_above_ema = pair_df["close"].iloc[-1] > pair_df["ema_long"].iloc[-1]
+                ema_short_above_ema_long = pair_df["ema_short"].iloc[-1] > pair_df["ema_long"].iloc[-1]
+                ema_short_slope = pair_df["ema_short_slope"].iloc[-1]
+
+                price_score = 0
+                if price_above_ema:
+                    price_score += 20
+                if ema_short_above_ema_long:
+                    price_score += 20
+                if ema_short_slope > 0.005:
+                    price_score += 15
+                elif ema_short_slope < -0.005:
+                    price_score -= 15
+
+                # K线形态
+                pair_df["bullish_engulfing"] = (
+                    (pair_df["close"].shift(1) < pair_df["open"].shift(1)) &
+                    (pair_df["close"] > pair_df["open"]) &
+                    (pair_df["close"] > pair_df["open"].shift(1)) &
+                    (pair_df["open"] < pair_df["close"].shift(1))
+                ).fillna(False)
+                pair_df["bearish_engulfing"] = (
+                    (pair_df["close"].shift(1) > pair_df["open"].shift(1)) &
+                    (pair_df["close"] < pair_df["open"]) &
+                    (pair_df["close"] < pair_df["open"].shift(1)) &
+                    (pair_df["open"] > pair_df["close"].shift(1))
+                ).fillna(False)
+
+                kline_score = 0
+                if pair_df["bullish_engulfing"].iloc[-1]:
+                    kline_score += 15
+                elif pair_df["bearish_engulfing"].iloc[-1]:
+                    kline_score -= 15
+                volatility = pair_df["close"].pct_change(10).std() * 100
+                if volatility > 0.5:
+                    kline_score += 10 if price_score > 0 else -10
+
+                # StochRSI
+                stochrsi = ta.STOCHRSI(pair_df, timeperiod=14, fastk_period=3, fastd_period=3)
+                pair_df["stochrsi_k"] = stochrsi["fastk"]
+                pair_df["stochrsi_d"] = stochrsi["fastd"]
+
+                stochrsi_score = 0
+                stochrsi_k = pair_df["stochrsi_k"].iloc[-1]
+                stochrsi_d = pair_df["stochrsi_d"].iloc[-1]
+                if stochrsi_k > 80 and stochrsi_k < stochrsi_d:
+                    stochrsi_score -= 15
+                elif stochrsi_k < 20 and stochrsi_k > stochrsi_d:
+                    stochrsi_score += 15
+                elif stochrsi_k > 50:
+                    stochrsi_score += 5
+                elif stochrsi_k < 50:
+                    stochrsi_score -= 5
+
+                # 量价关系
+                pair_df["volume_mean_20"] = pair_df["volume"].rolling(20).mean()
+                pair_df["volume_std_20"] = pair_df["volume"].rolling(20).std()
+                pair_df["volume_z_score"] = (pair_df["volume"] - pair_df["volume_mean_20"]) / pair_df["volume_std_20"]
+                pair_df["adx"] = ta.ADX(pair_df, timeperiod=14)
+
+                volume_score = 0
+                if pair_df["volume_z_score"].iloc[-1] > 1.5:
+                    volume_score += 10 if price_score > 0 else -10
+                if pair_df["adx"].iloc[-1] > 25:
+                    volume_score += 10 if price_score > 0 else -10
+
+                # 综合得分
+                raw_score = price_score + kline_score + stochrsi_score + volume_score
+                raw_score = max(min(raw_score, 50), -50)
+
+                # 对数映射到 [0, 100]
+                if raw_score >= 0:
+                    mapped_score = 50 + 50 * (np.log1p(raw_score / 50) / np.log1p(1))
+                else:
+                    mapped_score = 50 * (np.log1p(-raw_score / 50) / np.log1p(1))
+
+                trend_scores[tf] = max(0, min(100, int(round(mapped_score))))
+                logger.debug(f"[{pair}] {tf} 趋势得分：{trend_scores[tf]}, 原始得分：{raw_score}, "
+                            f"价格得分：{price_score}, K线得分：{kline_score}, "
+                            f"StochRSI得分：{stochrsi_score}, 量价得分：{volume_score}")
+
+            # 动态调整权重
+            if trend_scores.get("1h", 50) - trend_scores.get("3m", 50) > 20 or trend_scores.get("15m", 50) - trend_scores.get("3m", 50) > 20:
+                weights = {"3m": 0.2, "15m": 0.35, "1h": 0.45}
+                logger.debug(f"[{pair}] 1h 趋势得分（{trend_scores.get('1h', 50)}）显著高于 3m（{trend_scores.get('3m', 50)}），调整权重为 {weights}")
+
+            # 加权融合
+            final_score = sum(trend_scores[tf] * weights[tf] for tf in timeframes)
+            final_score = int(round(final_score))
+            final_score = max(0, min(100, final_score))
+
+            logger.info(f"[{pair}] 最终趋势得分：{final_score}, "
+                       f"3m得分：{trend_scores.get('3m', 50)}, 15m得分：{trend_scores.get('15m', 50)}, "
+                       f"1h得分：{trend_scores.get('1h', 50)}")
+            return final_score
+
+        except Exception as e:
+            logger.error(f"[{pair}] 获取市场趋势失败：{e}", exc_info=True)
+            return 50

freqtrade/templates/freqaiprimer.py.orig

@@ -0,0 +1,709 @@
+import logging
+import numpy as np
+import datetime
+import os
+import json
+import glob
+from functools import reduce
+from freqtrade.persistence import Trade
+import talib.abstract as ta
+from pandas import DataFrame
+from typing import Dict
+from freqtrade.strategy import (DecimalParameter, IStrategy, IntParameter)
+
+logger = logging.getLogger(__name__)
+
+class FreqaiPrimer(IStrategy):
+    """
+    基于 FreqAI 的动态阈值交易策略，集成动态加仓、减仓和自定义 ROI 逻辑，兼容最新 Freqtrade 版本
+    """
+
+    # --- 🧪 Hyperopt Parameters ---
+    TRAILING_STOP_START = DecimalParameter(0.01, 0.05, default=0.03, space='sell', optimize=True)
+    TRAILING_STOP_DISTANCE = DecimalParameter(0.005, 0.02, default=0.01, space='sell', optimize=True)
+
+    BUY_THRESHOLD_MIN = DecimalParameter(-0.1, -0.01, default=-0.05, space='buy', optimize=True)
+    BUY_THRESHOLD_MAX = DecimalParameter(-0.02, -0.001, default=-0.005, space='buy', optimize=True)
+    SELL_THRESHOLD_MIN = DecimalParameter(0.001, 0.02, default=0.005, space='sell', optimize=True)
+    SELL_THRESHOLD_MAX = DecimalParameter(0.02, 0.1, default=0.05, space='sell', optimize=True)
+
+    # 新增：加仓和减仓参数
+    ADD_POSITION_THRESHOLD = DecimalParameter(-0.05, -0.01, default=-0.02, space='buy', optimize=True)
+    EXIT_POSITION_RATIO = DecimalParameter(0.2, 0.7, default=0.5, space='sell', optimize=True)
+    COOLDOWN_PERIOD_MINUTES = IntParameter(1, 10, default=5, space='buy', optimize=True)
+    MAX_ENTRY_POSITION_ADJUSTMENT = IntParameter(1, 3, default=2, space='buy', optimize=True)
+
+    # --- 🛠️ 固定配置参数 ---
+    stoploss = -0.15
+    timeframe = "3m"
+    use_custom_stoploss = True
+    position_adjustment_enable = True  # 启用动态仓位调整
+
+    minimal_roi = {
+        "0": 0.08,    # 30分钟（0-30分钟）内，8% 盈利退出
+        "15": 0.04,   # 2小时（30-120分钟）内，4% 盈利退出
+        "60": 0.02,   # 4小时（120-240分钟）内，2% 盈利退出
+        "120": 0.0    # 8小时（240-480分钟）内，0% 盈利退出
+    }
+
+    plot_config = {
+        "main_plot": {
+            "ema200": {"color": "blue"},
+            "bb_upperband": {"color": "gray"},
+            "bb_lowerband": {"color": "gray"},
+            "bb_middleband": {"color": "gray"}
+        },
+        "subplots": {
+            "Signals": {
+                "enter_long": {"color": "green"},
+                "exit_long": {"color": "red"}
+            },
+            "Price-Value Divergence": {
+                "&-price_value_divergence": {"color": "purple"}
+            },
+            "Volume Z-Score": {
+                "volume_z_score": {"color": "orange"}
+            },
+            "RSI": {
+                "rsi": {"color": "cyan"}
+            }
+        }
+    }
+
+    freqai_info = {
+        "identifier": "test58",
+        "model": "LightGBMRegressor",
+        "feature_parameters": {
+            "include_timeframes": ["3m", "15m", "1h"],
+            "label_period_candles": 12,
+            "include_shifted_candles": 3,
+        },
+        "data_split_parameters": {
+            "test_size": 0.2,
+            "shuffle": False,
+        },
+        "model_training_parameters": {
+            "n_estimators": 200,
+            "learning_rate": 0.05,
+            "num_leaves": 31,
+            "verbose": -1,
+        },
+        "fit_live_predictions_candles": 100,
+        "live_retrain_candles": 100,
+    }
+
+    @staticmethod
+    def linear_map(value, from_min, from_max, to_min, to_max):
+        return (value - from_min) / (from_max - from_min) * (to_max - to_min) + to_min
+
+    def __init__(self, config: dict, *args, **kwargs):
+        super().__init__(config, *args, **kwargs)
+        logger.setLevel(logging.DEBUG)
+        logger.debug("✅ 策略已初始化，日志级别设置为 DEBUG")
+        self.trailing_stop_enabled = False
+        self.pair_stats = {}
+        self.stats_logged = False
+        self.fit_live_predictions_candles = self.freqai_info.get("fit_live_predictions_candles", 100)
+        self.last_entry_time = {}  # 记录每个币种的最后入场时间
+
+    def feature_engineering_expand_all(self, dataframe: DataFrame, period: int, metadata: dict, **kwargs) -> DataFrame:
+        dataframe["%-rsi-period"] = ta.RSI(dataframe, timeperiod=period)
+        dataframe["%-sma-period"] = ta.SMA(dataframe, timeperiod=period)
+        dataframe["%-ema-period"] = ta.EMA(dataframe, timeperiod=period)
+
+        real = ta.TYPPRICE(dataframe)
+        upperband, middleband, lowerband = ta.BBANDS(real, timeperiod=period, nbdevup=2.0, nbdevdn=2.0)
+        dataframe["bb_lowerband-period"] = lowerband
+        dataframe["bb_upperband-period"] = upperband
+        dataframe["bb_middleband-period"] = middleband
+        dataframe["%-bb_width-period"] = (dataframe["bb_upperband-period"] - dataframe["bb_lowerband-period"]) / dataframe["bb_middleband-period"]
+
+        dataframe["%-mfi-period"] = ta.MFI(dataframe, timeperiod=period)
+        dataframe["%-adx-period"] = ta.ADX(dataframe, timeperiod=period)
+        dataframe["%-relative_volume-period"] = dataframe["volume"] / dataframe["volume"].rolling(period).mean()
+
+        dataframe["ema200"] = ta.EMA(dataframe, timeperiod=200)
+        dataframe["%-price_value_divergence"] = (dataframe["close"] - dataframe["ema200"]) / dataframe["ema200"]
+
+        columns_to_clean = [
+            "%-rsi-period", "%-mfi-period", "%-sma-period", "%-ema-period", "%-adx-period",
+            "bb_lowerband-period", "bb_middleband-period", "bb_upperband-period",
+            "%-bb_width-period", "%-relative_volume-period", "%-price_value_divergence"
+        ]
+        for col in columns_to_clean:
+            dataframe[col] = dataframe[col].replace([np.inf, -np.inf], 0).ffill().fillna(0)
+
+        pair = metadata.get('pair', 'Unknown')
+        logger.debug(f"[{pair}] 特征工程完成，列：{list(dataframe.columns)}")
+        return dataframe
+
+    def set_freqai_targets(self, dataframe: DataFrame, metadata: dict, **kwargs) -> DataFrame:
+        pair = metadata.get('pair', 'Unknown')
+        if len(dataframe) < 200:
+            logger.warning(f"[{pair}] 数据量不足（{len(dataframe)}根K线），需要至少200根K线进行训练")
+            return dataframe
+
+        dataframe["ema200"] = ta.EMA(dataframe, timeperiod=200)
+        dataframe["&-price_value_divergence"] = (dataframe["close"] - dataframe["ema200"]) / dataframe["ema200"]
+
+        dataframe["volume_mean_20"] = dataframe["volume"].rolling(20).mean()
+        dataframe["volume_std_20"] = dataframe["volume"].rolling(20).std()
+        dataframe["volume_z_score"] = (dataframe["volume"] - dataframe["volume_mean_20"]) / dataframe["volume_std_20"]
+
+        dataframe["&-price_value_divergence"] = dataframe["&-price_value_divergence"].replace([np.inf, -np.inf], 0).ffill().fillna(0)
+        dataframe["volume_z_score"] = dataframe["volume_z_score"].replace([np.inf, -np.inf], 0).ffill().fillna(0)
+
+        return dataframe
+
+    def populate_indicators(self, dataframe: DataFrame, metadata: dict) -> DataFrame:
+        pair = metadata.get('pair', 'Unknown')
+        logger.info(f"[{pair}] 当前可用列（调用FreqAI前）：{list(dataframe.columns)}")
+
+        # 计算200周期EMA和历史价值背离
+        dataframe["ema200"] = ta.EMA(dataframe, timeperiod=200)
+        dataframe["price_value_divergence"] = (dataframe["close"] - dataframe["ema200"]) / dataframe["ema200"]
+
+        # 调用FreqAI预测价值背离
+        if not hasattr(self, 'freqai') or self.freqai is None:
+            logger.error(f"[{pair}] FreqAI 未初始化，请确保回测命令中启用了 --freqai")
+            dataframe["&-price_value_divergence"] = dataframe["price_value_divergence"]
+        else:
+            logger.debug(f"self.freqai 类型：{type(self.freqai)}")
+            dataframe = self.freqai.start(dataframe, metadata, self)
+
+            if "&-price_value_divergence" not in dataframe.columns:
+                logger.warning(f"[{pair}] 回归模型未生成 &-price_value_divergence，回退到规则计算")
+                dataframe["&-price_value_divergence"] = dataframe["price_value_divergence"]
+
+        # 计算其他指标
+        upperband, middleband, lowerband = ta.BBANDS(dataframe["close"], timeperiod=20, nbdevup=2.0, nbdevdn=2.0)
+        dataframe["bb_upperband"] = upperband
+        dataframe["bb_middleband"] = middleband
+        dataframe["bb_lowerband"] = lowerband
+        dataframe["rsi"] = ta.RSI(dataframe, timeperiod=14)
+        dataframe["volume_mean_20"] = dataframe["volume"].rolling(20).mean()
+        dataframe["volume_std_20"] = dataframe["volume"].rolling(20).std()
+        dataframe["volume_z_score"] = (dataframe["volume"] - dataframe["volume_mean_20"]) / dataframe["volume_std_20"]
+
+        # 数据清理
+        for col in ["ema200", "bb_upperband", "bb_middleband", "bb_lowerband", "rsi", "volume_z_score", "&-price_value_divergence", "price_value_divergence"]:
+            dataframe[col] = dataframe[col].replace([np.inf, -np.inf], 0).ffill().fillna(0)
+
+        # 添加调试日志
+        logger.debug(f"[{pair}] 最新数据 - close：{dataframe['close'].iloc[-1]:.6f}, "
+                    f"rsi：{dataframe['rsi'].iloc[-1]:.2f}, "
+                    f"&-price_value_divergence：{dataframe['&-price_value_divergence'].iloc[-1]:.6f}, "
+                    f"volume_z_score：{dataframe['volume_z_score'].iloc[-1]:.2f}, "
+                    f"bb_lowerband：{dataframe['bb_lowerband'].iloc[-1]:.6f}")
+
+        # 获取 labels_mean 和 labels_std
+        labels_mean = None
+        labels_std = None
+
+        try:
+            model_base_dir = os.path.join(self.config["user_data_dir"], "models", self.freqai_info["identifier"])
+            pair_base = pair.split('/')[0] if '/' in pair else pair
+            sub_dirs = glob.glob(os.path.join(model_base_dir, f"sub-train-{pair_base}_*"))
+
+            if not sub_dirs:
+                logger.warning(f"[{pair}] 未找到任何子目录：{model_base_dir}/sub-train-{pair_base}_*")
+            else:
+                latest_sub_dir = max(sub_dirs, key=lambda x: int(x.split('_')[-1]))
+                pair_base_lower = pair_base.lower()
+                timestamp = latest_sub_dir.split('_')[-1]
+                metadata_file = os.path.join(latest_sub_dir, f"cb_{pair_base_lower}_{timestamp}_metadata.json")
+
+                if os.path.exists(metadata_file):
+                    with open(metadata_file, "r") as f:
+                        metadata = json.load(f)
+                        labels_mean = metadata["labels_mean"]["&-price_value_divergence"]
+                        labels_std = metadata["labels_std"]["&-price_value_divergence"]
+                        logger.info(f"[{pair}] 从最新子目录 {latest_sub_dir} 读取 labels_mean：{labels_mean}, labels_std：{labels_std}")
+                else:
+                    logger.warning(f"[{pair}] 最新的 metadata.json 文件 {metadata_file} 不存在")
+        except Exception as e:
+            logger.warning(f"[{pair}] 无法从子目录读取 labels_mean 和 labels_std：{e}，重新计算")
+
+        if labels_mean is None or labels_std is None:
+            logger.warning(f"[{pair}] 无法获取 labels_mean 和 labels_std，重新计算")
+            dataframe["&-price_value_divergence_actual"] = (dataframe["close"] - dataframe["ema200"]) / dataframe["ema200"]
+            dataframe["&-price_value_divergence_actual"] = dataframe["&-price_value_divergence_actual"].replace([np.inf, -np.inf], 0).ffill().fillna(0)
+            recent_data = dataframe["&-price_value_divergence_actual"].tail(self.fit_live_predictions_candles)
+            labels_mean = recent_data.mean()
+            labels_std = recent_data.std()
+            if np.isnan(labels_std) or labels_std == 0:
+                labels_std = 0.01
+                logger.warning(f"[{pair}] labels_std 计算异常，使用默认值 0.01")
+
+        # 根据市场趋势得分动态调整买卖阈值
+        market_trend_score = self.get_market_trend(dataframe=dataframe, metadata={'pair': pair})
+        k_buy = FreqaiPrimer.linear_map(market_trend_score, 0, 100, 1.2, 0.8)
+        k_sell = FreqaiPrimer.linear_map(market_trend_score, 0, 100, 1.5, 1.0)
+
+        self.buy_threshold = labels_mean - k_buy * labels_std
+        self.sell_threshold = labels_mean + k_sell * labels_std
+
+        # 使用 Hyperopt 参数限制阈值
+        self.buy_threshold = max(self.buy_threshold, self.BUY_THRESHOLD_MIN.value)
+        self.buy_threshold = min(self.buy_threshold, self.BUY_THRESHOLD_MAX.value)
+        self.sell_threshold = min(self.sell_threshold, self.SELL_THRESHOLD_MAX.value)
+        self.sell_threshold = max(self.sell_threshold, self.SELL_THRESHOLD_MIN.value)
+
+        logger.info(f"[{pair}] 市场趋势得分：{market_trend_score}, labels_mean：{labels_mean:.4f}, labels_std：{labels_std:.4f}")
+        logger.info(f"[{pair}] k_buy：{k_buy:.2f}, k_sell：{k_sell:.2f}")
+        logger.info(f"[{pair}] 动态买入阈值：{self.buy_threshold:.4f}, 卖出阈值：{self.sell_threshold:.4f}")
+
+        if not self.stats_logged:
+            logger.info("===== 所有币对的 labels_mean 和 labels_std 汇总 =====")
+            for p, stats in self.pair_stats.items():
+                logger.info(f"[{p}] labels_mean：{stats['labels_mean']:.4f}, labels_std：{stats['labels_std']:.4f}")
+            logger.info("==============================================")
+            self.stats_logged = True
+
+        return dataframe
+
+    def trailing_space(self):
+        return [
+            DecimalParameter(0.01, 0.05, name="trailing_stop_start"),
+            DecimalParameter(0.005, 0.02, name="trailing_stop_distance")
+        ]
+
+    def leverage_space(self):
+        return [
+            DecimalParameter(-0.05, -0.01, name="add_position_threshold", default=-0.02),
+            DecimalParameter(0.2, 0.7, name="exit_position_ratio", default=0.5),
+            IntParameter(1, 10, name="cooldown_period_minutes", default=5),
+            IntParameter(1, 3, name="max_entry_position_adjustment", default=2)
+        ]
+
+    def populate_entry_trend(self, dataframe: DataFrame, metadata: dict) -> DataFrame:
+        pair = metadata.get('pair', 'Unknown')
+        conditions = []
+
+        # 获取市场趋势得分
+        trend_score = self.get_market_trend(dataframe=dataframe, metadata=metadata)
+
+        # 动态调整成交量阈值：牛市（trend_score=100）-> 0.5，熊市（trend_score=0）-> 1.5
+        volume_z_score_min = 0.5
+        volume_z_score_max = 1.5
+        volume_z_score_threshold = self.linear_map(trend_score, 0, 100, volume_z_score_max, volume_z_score_min)
+
+        # 动态调整 RSI 阈值：牛市（trend_score=100）-> 40，熊市（trend_score=0）-> 60
+        rsi_min = 40
+        rsi_max = 60
+        rsi_threshold = self.linear_map(trend_score, 0, 100, rsi_max, rsi_min)
+
+        # 新增：动态调整 STOCHRSI 阈值，牛市 -> 30，熊市 -> 50
+        stochrsi_min = 30
+        stochrsi_max = 50
+        stochrsi_threshold = self.linear_map(trend_score, 0, 100, stochrsi_max, stochrsi_min)
+
+        if "&-price_value_divergence" in dataframe.columns:
+            # 计算 STOCHRSI
+            stochrsi = ta.STOCHRSI(dataframe, timeperiod=14, fastk_period=3, fastd_period=3)
+            dataframe["stochrsi_k"] = stochrsi["fastk"]
+
+            cond1 = (dataframe["&-price_value_divergence"] < self.buy_threshold)
+            cond2 = (dataframe["volume_z_score"] > volume_z_score_threshold)
+            cond3 = (dataframe["rsi"] < rsi_threshold)
+            cond4 = (dataframe["close"] <= dataframe["bb_lowerband"])
+            cond5 = (dataframe["stochrsi_k"] < stochrsi_threshold)  # 新增 STOCHRSI 条件
+            buy_condition = cond1 & cond2 & cond3 & cond4 & cond5
+            conditions.append(buy_condition)
+
+            divergence_value = dataframe['&-price_value_divergence'].iloc[-1] if not dataframe['&-price_value_divergence'].isna().all() else np.nan
+            volume_z_score_value = dataframe['volume_z_score'].iloc[-1] if not dataframe['volume_z_score'].isna().all() else np.nan
+            rsi_value = dataframe['rsi'].iloc[-1] if not dataframe['rsi'].isna().all() else np.nan
+            stochrsi_value = dataframe['stochrsi_k'].iloc[-1] if not dataframe['stochrsi_k'].isna().all() else np.nan
+            logger.debug(f"[{pair}] 买入条件检查 - "
+                        f"&-price_value_divergence={divergence_value:.6f} < {self.buy_threshold:.6f}: {cond1.iloc[-1]}, "
+                        f"volume_z_score={volume_z_score_value:.2f} > {volume_z_score_threshold:.2f}: {cond2.iloc[-1]}, "
+                        f"rsi={rsi_value:.2f} < {rsi_threshold:.2f}: {cond3.iloc[-1]}, "
+                        f"close={dataframe['close'].iloc[-1]:.6f} <= bb_lowerband={dataframe['bb_lowerband'].iloc[-1]:.6f}: {cond4.iloc[-1]}, "
+                        f"stochrsi_k={stochrsi_value:.2f} < {stochrsi_threshold:.2f}: {cond5.iloc[-1]}")
+        else:
+            logger.warning(f"[{pair}] ⚠️ &-price_value_divergence 列缺失，跳过买入信号生成")
+
+        if len(conditions) > 0:
+            combined_condition = reduce(lambda x, y: x & y, conditions)
+            if combined_condition.any():
+                dataframe.loc[combined_condition, 'enter_long'] = 1
+                logger.info(f"[{pair}] 买入信号触发，条件满足，趋势得分：{trend_score:.2f}")
+            else:
+                logger.debug(f"[{pair}] 买入条件未满足，无买入信号")
+        else:
+            logger.debug(f"[{pair}] 无有效买入条件")
+
+        return dataframe
+
+    def populate_exit_trend(self, dataframe: DataFrame, metadata: dict) -> DataFrame:
+        pair = metadata.get('pair', 'Unknown')
+        conditions = []
+
+        if "&-price_value_divergence" in dataframe.columns:
+            cond1 = (dataframe["&-price_value_divergence"] > self.sell_threshold * 1.025)
+            cond2 = (dataframe["rsi"] > 75)
+            sell_condition = cond1 | cond2
+            conditions.append(sell_condition)
+
+            logger.debug(f"[{pair}] 卖出条件检查 - "
+                         f"&-price_value_divergence > {self.sell_threshold:.6f}: {cond1.iloc[-1]}, "
+                         f"rsi > 75: {cond2.iloc[-1]}")
+        else:
+            logger.warning(f"[{pair}] ⚠️ &-price_value_divergence 列缺失，跳过该条件")
+
+        if len(conditions) > 0:
+            dataframe.loc[reduce(lambda x, y: x & y, conditions), 'exit_long'] = 1
+            logger.debug(f"[{pair}] 出场信号触发，条件满足")
+        else:
+            logger.debug(f"[{pair}] 无有效卖出条件")
+
+        return dataframe
+
+    def buy_space(self):
+        return [
+            DecimalParameter(-0.1, -0.01, name="buy_threshold_min"),
+            DecimalParameter(-0.02, -0.001, name="buy_threshold_max"),
+            DecimalParameter(-0.05, -0.01, name="add_position_threshold", default=-0.02),
+            IntParameter(1, 10, name="cooldown_period_minutes", default=5),
+            IntParameter(1, 3, name="max_entry_position_adjustment", default=2)
+        ]
+
+    def sell_space(self):
+        return [
+            DecimalParameter(0.001, 0.02, name="sell_threshold_min"),
+            DecimalParameter(0.02, 0.1, name="sell_threshold_max"),
+            DecimalParameter(0.2, 0.7, name="exit_position_ratio", default=0.5)
+        ]
+
+    def adjust_trade_position(self, trade: Trade, current_time: datetime,
+                            current_rate: float, current_profit: float,
+                            min_stake: float | None, max_stake: float,
+                            current_entry_rate: float, current_exit_rate: float,
+                            current_entry_profit: float, current_exit_profit: float,
+                            **kwargs) -> float | None | tuple[float | None, str | None]:
+        """
+        动态调整仓位：支持加仓、减仓、追踪止损和最大持仓时间限制
+        参数：
+        - trade: 当前交易对象
+        - current_time: 当前时间
+        - current_rate: 当前价格
+        - current_profit: 当前总盈利
+        - min_stake: 最小下注金额
+        - max_stake: 最大下注金额
+        - current_entry_rate: 当前入场价格
+        - current_exit_rate: 当前退出价格
+        - current_entry_profit: 当前入场盈利
+        - current_exit_profit: 当前退出盈利
+        返回：
+        - 调整金额（正数为加仓，负数为减仓）或 None
+        """
+        pair = trade.pair
+        dataframe = self.dp.get_pair_dataframe(pair, self.timeframe)
+        trend_score = self.get_market_trend(dataframe=dataframe, metadata={'pair': pair})
+        hold_time = (current_time - trade.open_date_utc).total_seconds() / 60
+        profit_ratio = (current_rate - trade.open_rate) / trade.open_rate
+
+        initial_stake_amount = trade.stake_amount / 3
+        logger.debug(f"{pair} 首次入场金额: {initial_stake_amount:.2f}, 当前持仓金额: {trade.stake_amount:.2f}, "
+                    f"加仓次数: {trade.nr_of_successful_entries - 1}")
+
+        # 加仓逻辑
+        max_entry_adjustments = self.MAX_ENTRY_POSITION_ADJUSTMENT.value
+        if trade.nr_of_successful_entries <= max_entry_adjustments + 1:
+            add_position_threshold = self.ADD_POSITION_THRESHOLD.value
+            # 线性映射加仓阈值，趋势值越高，加仓越严格
+            add_threshold = 80 - 30 * (trend_score / 100)  # 趋势值 100 -> 50, 0 -> 80
+            if profit_ratio <= add_position_threshold and hold_time > 5 and trend_score <= add_threshold:
+                logger.debug(f"{pair} 初始下注金额: {initial_stake_amount:.2f}, trend_score: {trend_score:.2f}, add_threshold: {add_threshold} ")
+
+                # 计算加仓金额
+                add_count = trade.nr_of_successful_entries - 1
+                multipliers = [2, 4, 8]
+                if add_count < len(multipliers):
+                    multiplier = multipliers[add_count]
+                    add_amount = initial_stake_amount * multiplier
+                    logger.debug(f"{pair} 第 {add_count + 1} 次加仓，倍数={multiplier}, "
+                                f"金额 = {initial_stake_amount:.2f} * {multiplier} = {add_amount:.2f}")
+                    logger.debug(f"{pair} 加仓计算: 第 {add_count + 1} 次加仓，倍数={multiplier}, "
+                                f"金额 = {initial_stake_amount:.2f} * {multiplier} = {add_amount:.2f}")
+
+                    if min_stake is not None and add_amount < min_stake:
+                        logger.warning(f"{pair} 加仓金额 {add_amount:.2f} 低于最小下注金额 {min_stake:.2f}，取消加仓")
+                        return (None, f"Add amount {add_amount:.2f} below min_stake {min_stake:.2f}")
+                    if add_amount > max_stake:
+                        logger.warning(f"{pair} 加仓金额 {add_amount:.2f} 超出最大可用金额 {max_stake:.2f}，调整为 {max_stake:.2f}")
+                        add_amount = max_stake
+                    logger.info(f"{pair} 价格下跌 {profit_ratio*100:.2f}%，触发第 {add_count + 1} 次加仓 {add_amount:.2f}")
+                    return (add_amount, f"Price dropped {profit_ratio*100:.2f}%, add {add_amount:.2f}")
+
+        # 减仓逻辑
+        exit_position_ratio = self.EXIT_POSITION_RATIO.value
+        if profit_ratio >= 0.03:
+            # 趋势值越高，减仓比例越低
+            reduce_factor = 0.6 + 0.4 * (1 - trend_score / 100)  # 牛市(100) -> 0.6, 熊市(0) -> 1.0
+            reduce_amount = -exit_position_ratio * reduce_factor * trade.stake_amount
+            logger.info(f"{pair} 趋势值 {trend_score:.2f}，利润 {profit_ratio*100:.2f}%，减仓 {abs(reduce_amount):.2f}")
+            return (reduce_amount, f"Profit {profit_ratio*100:.2f}%")
+        elif profit_ratio >= 0.05:
+            reduce_factor = 1.4 - 0.4 * (trend_score / 100)  # 牛市(100) -> 1.0, 熊市(0) -> 1.4
+            reduce_amount = -exit_position_ratio * reduce_factor * trade.stake_amount
+            logger.info(f"{pair} 趋势值 {trend_score:.2f}，利润 {profit_ratio*100:.2f}%，减仓 {abs(reduce_amount):.2f}")
+            return (reduce_amount, f"Profit {profit_ratio*100:.2f}%")
+
+        # 追踪止损逻辑
+        trailing_stop_start = self.TRAILING_STOP_START.value
+        trailing_stop_distance = self.TRAILING_STOP_DISTANCE.value
+        # 使用 Sigmoid 映射调整追踪止损参数
+        sigmoid = 1 / (1 + np.exp(-0.1 * (trend_score - 50)))
+        trailing_factor = 0.8 + (1.2 - 0.8) * sigmoid  # 牛市(100) -> 1.2, 熊市(0) -> 0.8
+        distance_factor = 0.7 + (1.5 - 0.7) * sigmoid  # 牛市(100) -> 1.5, 熊市(0) -> 0.7
+        trailing_stop_start *= trailing_factor
+        trailing_stop_distance *= distance_factor
+
+        if profit_ratio >= trailing_stop_start and not self.trailing_stop_enabled:
+            self.trailing_stop_enabled = True
+            trade.adjust_min_max_rates(current_rate, current_rate)
+            logger.info(f"{pair} 价格上涨超过 {trailing_stop_start*100:.1f}%，启动追踪止损")
+            return None
+
+        if self.trailing_stop_enabled:
+            max_rate = trade.max_rate or current_rate
+            trailing_stop_price = max_rate * (1 - trailing_stop_distance)
+            if current_rate < trailing_stop_price:
+                logger.info(f"{pair} 价格回落至 {trailing_stop_price:.6f}，触发全部卖出")
+                return (-trade.stake_amount, f"Trailing stop at {trailing_stop_price:.6f}")
+            trade.adjust_min_max_rates(current_rate, trade.min_rate)
+            return None
+
+        return None
+
+    def confirm_trade_entry(self, pair: str, order_type: str, amount: float, rate: float,
+                            time_in_force: str, current_time: datetime, **kwargs) -> bool:
+        # 调试日志：记录输入参数
+        logger.debug(f"[{pair}] confirm_trade_entry called with rate={rate}, type(rate)={type(rate)}, "
+                    f"amount={amount}, order_type={order_type}, time_in_force={time_in_force}")
+
+        # 检查 rate 是否有效
+        if not isinstance(rate, (float, int)) or rate is None:
+            logger.error(f"[{pair}] Invalid rate value: {rate} (type: {type(rate)}). Skipping trade entry.")
+            return False
+
+        market_trend_score = self.get_market_trend(dataframe=DataFrame, metadata={'pair': pair})
+        cooldown_period_minutes = self.COOLDOWN_PERIOD_MINUTES.value if market_trend_score > 50 else self.COOLDOWN_PERIOD_MINUTES.value // 2
+
+        if pair in self.last_entry_time:
+            last_time = self.last_entry_time[pair]
+            if (current_time - last_time).total_seconds() < cooldown_period_minutes * 60:
+                logger.info(f"[{pair}] 冷却期内（{cooldown_period_minutes} 分钟），跳过本次入场")
+                return False
+
+        self.last_entry_time[pair] = current_time
+        self.trailing_stop_enabled = False
+        try:
+            logger.info(f"[{pair}] 确认入场，价格：{float(rate):.6f}")
+        except (ValueError, TypeError) as e:
+            logger.error(f"[{pair}] Failed to format rate: {rate} (type: {type(rate)}), error: {e}")
+            return False
+
+        return True
+
+    def confirm_trade_exit(self, pair: str, trade: Trade, order_type: str, amount: float,
+                           rate: float, time_in_force: str, exit_reason: str,
+                           current_time: datetime, **kwargs) -> bool:
+        adjusted_rate = rate * (1 + 0.0025)
+        logger.info(f"[{pair}] 退出交易，原因：{exit_reason}, 原始利润：{trade.calc_profit_ratio(rate):.2%},"
+                    f"调整后卖出价：{adjusted_rate:.6f}")
+        return True
+
+    def custom_roi(self, trade: Trade, current_profit: float, current_time: datetime, trade_dur: int, 
+                   current_rate: float = None, min_stake: float | None = None, max_stake: float | None = None) -> dict:
+        """
+        动态调整 ROI 表格，基于 FreqAI 预测的 &-price_value_divergence 和 RSI。
+        - 负的 divergence（预测上涨）或低 RSI 时提高 ROI。
+        - 正的 divergence（预测下跌）或高 RSI 时降低 ROI。
+        - 长时间持仓降低 ROI 目标。
+        """
+        pair = trade.pair
+        logger.debug(f"[{pair}] 计算自定义 ROI，当前盈利: {current_profit:.2%}, 持仓时间: {trade_dur} 分钟")
+
+        # 获取最新数据
+        dataframe = self.dp.get_pair_dataframe(pair=pair, timeframe=self.timeframe)
+        dataframe = self.populate_indicators(dataframe, {'pair': pair})  # 计算指标
+
+        # 获取 FreqAI 预测和 RSI
+        divergence = dataframe["&-price_value_divergence"].iloc[-1] if "&-price_value_divergence" in dataframe else 0
+        rsi = dataframe["rsi"].iloc[-1] if "rsi" in dataframe else 50
+
+        # 计算调整系数
+        # 1. Divergence 调整：负值（预测上涨）-> 提高 ROI，正值（预测下跌）-> 降低 ROI
+        divergence_factor = self.linear_map(divergence, -0.1, 0.1, 1.2, 0.8)
+        
+        # 2. RSI 调整：低 RSI（超卖）-> 提高 ROI，高 RSI（超买）-> 降低 ROI
+        rsi_factor = self.linear_map(rsi, 30, 70, 1.2, 0.8)
+        
+        # 3. 时间调整：持仓时间越长，ROI 目标降低
+        time_factor = self.linear_map(trade_dur, 0, 240, 1.0, 0.7)  # 4小时后 ROI 降低到 70%
+
+        # 综合调整系数
+        roi_factor = divergence_factor * rsi_factor * time_factor
+
+        # 默认 ROI 表格
+        base_roi = {
+            0: 0.08,
+            15: 0.04,
+            60: 0.02,
+            120: 0.0
+        }
+
+        # 动态调整 ROI，限制在 0% 到 20% 之间
+        dynamic_roi = {time: min(max(roi * roi_factor, 0.0), 0.2) for time, roi in base_roi.items()}
+        
+        logger.debug(f"[{pair}] Divergence: {divergence:.4f}, RSI: {rsi:.2f}, 持仓时间: {trade_dur} 分钟, "
+                     f"调整系数: divergence={divergence_factor:.2f}, rsi={rsi_factor:.2f}, time={time_factor:.2f}, "
+                     f"总系数={roi_factor:.2f}, 动态 ROI 表格: {dynamic_roi}")
+        return dynamic_roi
+
+    def custom_entry_price(self, pair: str, trade: Trade | None, current_time: datetime, proposed_rate: float,
+                           entry_tag: str | None, side: str, **kwargs) -> float:
+        adjusted_rate = proposed_rate * (1 - 0.005)
+        logger.debug(f"[{pair}] 自定义买入价：{adjusted_rate:.6f}（原价：{proposed_rate:.6f}）")
+        return adjusted_rate
+
+    def custom_exit_price(self, pair: str, trade: Trade,
+                          current_time: datetime, proposed_rate: float,
+                          current_profit: float, exit_tag: str | None, **kwargs) -> float:
+        adjusted_rate = proposed_rate * (1 + 0.0025)
+        logger.debug(f"[{pair}] 自定义卖出价：{adjusted_rate:.6f}（原价：{proposed_rate:.6f}）")
+        return adjusted_rate
+
+    def get_market_trend(self, dataframe: DataFrame = None, metadata: dict = None) -> int:
+        try:
+            timeframes = ["3m", "15m", "1h"]
+            weights = {"3m": 0.5, "15m": 0.3, "1h": 0.2}
+            trend_scores = {}
+            pair = metadata.get('pair', 'Unknown') if metadata else 'Unknown'
+
+            # 检查 pair 是否有效
+            if pair == 'Unknown':
+                logger.error(f"[{pair}] Invalid pair in metadata: {metadata}. Returning default score 50")
+                return 50
+
+            logger.debug(f"[{pair}] 正在计算多时间框架市场趋势得分")
+
+            for tf in timeframes:
+                # 优先使用传入的 dataframe（如果匹配主时间框架），否则加载目标币对数据
+                pair_df = dataframe if tf == self.timeframe and dataframe is not None else self.dp.get_pair_dataframe(pair=pair, timeframe=tf)
+
+                min_candles = 200 if tf == "3m" else 100 if tf == "15m" else 50
+                if pair_df.empty or len(pair_df) < min_candles:
+                    logger.warning(f"[{pair}] 数据不足（{tf}），使用默认得分：50")
+                    trend_scores[tf] = 50
+                    continue
+
+                # 价格趋势
+                ema_short_period = 50 if tf == "3m" else 20 if tf == "15m" else 12
+                ema_long_period = 200 if tf == "3m" else 80 if tf == "15m" else 50
+                pair_df["ema_short"] = ta.EMA(pair_df, timeperiod=ema_short_period)
+                pair_df["ema_long"] = ta.EMA(pair_df, timeperiod=ema_long_period)
+                pair_df["ema_short_slope"] = (pair_df["ema_short"] - pair_df["ema_short"].shift(10)) / pair_df["ema_short"].shift(10)
+
+                price_above_ema = pair_df["close"].iloc[-1] > pair_df["ema_long"].iloc[-1]
+                ema_short_above_ema_long = pair_df["ema_short"].iloc[-1] > pair_df["ema_long"].iloc[-1]
+                ema_short_slope = pair_df["ema_short_slope"].iloc[-1]
+
+                price_score = 0
+                if price_above_ema:
+                    price_score += 20
+                if ema_short_above_ema_long:
+                    price_score += 20
+                if ema_short_slope > 0.005:
+                    price_score += 15
+                elif ema_short_slope < -0.005:
+                    price_score -= 15
+
+                # K线形态
+                pair_df["bullish_engulfing"] = (
+                    (pair_df["close"].shift(1) < pair_df["open"].shift(1)) &
+                    (pair_df["close"] > pair_df["open"]) &
+                    (pair_df["close"] > pair_df["open"].shift(1)) &
+                    (pair_df["open"] < pair_df["close"].shift(1))
+                ).fillna(False)
+                pair_df["bearish_engulfing"] = (
+                    (pair_df["close"].shift(1) > pair_df["open"].shift(1)) &
+                    (pair_df["close"] < pair_df["open"]) &
+                    (pair_df["close"] < pair_df["open"].shift(1)) &
+                    (pair_df["open"] > pair_df["close"].shift(1))
+                ).fillna(False)
+
+                kline_score = 0
+                if pair_df["bullish_engulfing"].iloc[-1]:
+                    kline_score += 15
+                elif pair_df["bearish_engulfing"].iloc[-1]:
+                    kline_score -= 15
+                volatility = pair_df["close"].pct_change(10).std() * 100
+                if volatility > 0.5:
+                    kline_score += 10 if price_score > 0 else -10
+
+                # StochRSI
+                stochrsi = ta.STOCHRSI(pair_df, timeperiod=14, fastk_period=3, fastd_period=3)
+                pair_df["stochrsi_k"] = stochrsi["fastk"]
+                pair_df["stochrsi_d"] = stochrsi["fastd"]
+
+                stochrsi_score = 0
+                stochrsi_k = pair_df["stochrsi_k"].iloc[-1]
+                stochrsi_d = pair_df["stochrsi_d"].iloc[-1]
+                if stochrsi_k > 80 and stochrsi_k < stochrsi_d:
+                    stochrsi_score -= 15
+                elif stochrsi_k < 20 and stochrsi_k > stochrsi_d:
+                    stochrsi_score += 15
+                elif stochrsi_k > 50:
+                    stochrsi_score += 5
+                elif stochrsi_k < 50:
+                    stochrsi_score -= 5
+
+                # 量价关系
+                pair_df["volume_mean_20"] = pair_df["volume"].rolling(20).mean()
+                pair_df["volume_std_20"] = pair_df["volume"].rolling(20).std()
+                pair_df["volume_z_score"] = (pair_df["volume"] - pair_df["volume_mean_20"]) / pair_df["volume_std_20"]
+                pair_df["adx"] = ta.ADX(pair_df, timeperiod=14)
+
+                volume_score = 0
+                if pair_df["volume_z_score"].iloc[-1] > 1.5:
+                    volume_score += 10 if price_score > 0 else -10
+                if pair_df["adx"].iloc[-1] > 25:
+                    volume_score += 10 if price_score > 0 else -10
+
+                # 综合得分
+                raw_score = price_score + kline_score + stochrsi_score + volume_score
+                raw_score = max(min(raw_score, 50), -50)
+
+                # 对数映射到 [0, 100]
+                if raw_score >= 0:
+                    mapped_score = 50 + 50 * (np.log1p(raw_score / 50) / np.log1p(1))
+                else:
+                    mapped_score = 50 * (np.log1p(-raw_score / 50) / np.log1p(1))
+
+                trend_scores[tf] = max(0, min(100, int(round(mapped_score))))
+                logger.debug(f"[{pair}] {tf} 趋势得分：{trend_scores[tf]}, 原始得分：{raw_score}, "
+                            f"价格得分：{price_score}, K线得分：{kline_score}, "
+                            f"StochRSI得分：{stochrsi_score}, 量价得分：{volume_score}")
+
+            # 动态调整权重
+            if trend_scores.get("1h", 50) - trend_scores.get("3m", 50) > 20:
+                weights = {"3m": 0.3, "15m": 0.3, "1h": 0.4}
+                logger.debug(f"[{pair}] 1h 趋势得分（{trend_scores.get('1h', 50)}）显著高于 3m（{trend_scores.get('3m', 50)}），调整权重为 {weights}")
+
+            # 加权融合
+            final_score = sum(trend_scores[tf] * weights[tf] for tf in timeframes)
+            final_score = int(round(final_score))
+            final_score = max(0, min(100, final_score))
+
+            logger.info(f"[{pair}] 最终趋势得分：{final_score}, "
+                       f"3m得分：{trend_scores.get('3m', 50)}, 15m得分：{trend_scores.get('15m', 50)}, "
+                       f"1h得分：{trend_scores.get('1h', 50)}")
+            return final_score
+
+        except Exception as e:
+            logger.error(f"[{pair}] 获取市场趋势失败：{e}", exc_info=True)
+            return 50

freqtrade/templates/freqprimer.py_origin

@@ -8,27 +8,33 @@ from freqtrade.strategy import IStrategy, IntParameter, DecimalParameter
 
 logger = logging.getLogger(__name__)
 
-class FreqaiExampleStrategy(IStrategy):
-    # 移除硬编码的 minimal_roi 和 stoploss，改为动态适配
-    minimal_roi = {}  # 将在 populate_indicators 中动态生成
-    stoploss = 0.0   # 将在 populate_indicators 中动态设置
+class FreqaiPrimer(IStrategy):
+    minimal_roi = {
+        0: 0.135,
+        9: 0.052,
+        15: 0.007,
+        60: 0
+    }
+    stoploss = -0.263
     trailing_stop = True
+    trailing_stop_positive = 0.324
+    trailing_stop_positive_offset = 0.411
+    trailing_only_offset_is_reached = False
+    max_open_trades = 4
     process_only_new_candles = True
     use_exit_signal = True
     startup_candle_count: int = 40
     can_short = False
 
-    # 参数定义：FreqAI 动态适配 buy_rsi 和 sell_rsi，禁用 Hyperopt 优化
-    buy_rsi = IntParameter(low=10, high=50, default=27, space="buy", optimize=False, load=True)
-    sell_rsi = IntParameter(low=50, high=90, default=59, space="sell", optimize=False, load=True)
+    buy_rsi = IntParameter(low=10, high=50, default=30, space="buy", optimize=False, load=True)
+    sell_rsi = IntParameter(low=50, high=90, default=70, space="sell", optimize=False, load=True)
+    roi_0 = DecimalParameter(low=0.01, high=0.2, default=0.135, space="roi", optimize=True, load=True)
+    roi_15 = DecimalParameter(low=0.005, high=0.1, default=0.052, space="roi", optimize=True, load=True)
+    roi_30 = DecimalParameter(low=0.001, high=0.05, default=0.007, space="roi", optimize=True, load=True)
+    stoploss_param = DecimalParameter(low=-0.35, high=-0.1, default=-0.263, space="stoploss", optimize=True, load=True)
+    trailing_stop_positive_param = DecimalParameter(low=0.1, high=0.5, default=0.324, space="trailing", optimize=True, load=True)
+    trailing_stop_positive_offset_param = DecimalParameter(low=0.2, high=0.6, default=0.411, space="trailing", optimize=True, load=True)
 
-    # 为 Hyperopt 优化添加 ROI 和 stoploss 参数
-    roi_0 = DecimalParameter(low=0.01, high=0.2, default=0.038, space="roi", optimize=True, load=True)
-    roi_15 = DecimalParameter(low=0.005, high=0.1, default=0.027, space="roi", optimize=True, load=True)
-    roi_30 = DecimalParameter(low=0.001, high=0.05, default=0.009, space="roi", optimize=True, load=True)
-    stoploss_param = DecimalParameter(low=-0.35, high=-0.1, default=-0.182, space="stoploss", optimize=True, load=True)
-
-    # FreqAI 配置
     freqai_info = {
         "model": "LightGBMRegressor",
         "feature_parameters": {
@@ -42,8 +48,8 @@ class FreqaiExampleStrategy(IStrategy):
             "shuffle": False,
         },
         "model_training_parameters": {
-            "n_estimators": 100,
-            "learning_rate": 0.1,
+            "n_estimators": 200,
+            "learning_rate": 0.05,
             "num_leaves": 31,
             "verbose": -1,
         },
@@ -78,26 +84,29 @@ class FreqaiExampleStrategy(IStrategy):
         dataframe["%-relative_volume-period"] = (
             dataframe["volume"] / dataframe["volume"].rolling(period).mean()
         )
-        dataframe.replace([np.inf, -np.inf], 0, inplace=True)
-        dataframe.ffill(inplace=True)
-        dataframe.fillna(0, inplace=True)
+        dataframe = dataframe.replace([np.inf, -np.inf], 0)
+        dataframe = dataframe.ffill()
+        dataframe = dataframe.fillna(0)
         return dataframe
 
     def feature_engineering_expand_basic(self, dataframe: DataFrame, metadata: dict, **kwargs) -> DataFrame:
         dataframe["%-pct-change"] = dataframe["close"].pct_change()
         dataframe["%-raw_volume"] = dataframe["volume"]
         dataframe["%-raw_price"] = dataframe["close"]
-        dataframe.replace([np.inf, -np.inf], 0, inplace=True)
-        dataframe.fillna(method='ffill', inplace=True)
-        dataframe.fillna(0, inplace=True)
+        dataframe = dataframe.replace([np.inf, -np.inf], 0)
+        dataframe = dataframe.ffill()
+        dataframe = dataframe.fillna(0)
         return dataframe
 
     def feature_engineering_standard(self, dataframe: DataFrame, metadata: dict, **kwargs) -> DataFrame:
+        if len(dataframe["close"]) < 20:
+            logger.warning(f"数据不足 {len(dataframe)} 根 K 线，%-volatility 可能不完整")
         dataframe["%-day_of_week"] = dataframe["date"].dt.dayofweek
         dataframe["%-hour_of_day"] = dataframe["date"].dt.hour
-        dataframe.replace([np.inf, -np.inf], 0, inplace=True)
-        dataframe.fillna(method='ffill', inplace=True)
-        dataframe.fillna(0, inplace=True)
+        dataframe["%-volatility"] = dataframe["close"].pct_change().rolling(20, min_periods=1).std()
+        dataframe["%-volatility"] = dataframe["%-volatility"].replace([np.inf, -np.inf], 0)
+        dataframe["%-volatility"] = dataframe["%-volatility"].ffill()
+        dataframe["%-volatility"] = dataframe["%-volatility"].fillna(0)
         return dataframe
 
     def set_freqai_targets(self, dataframe: DataFrame, metadata: dict, **kwargs) -> DataFrame:
@@ -105,84 +114,83 @@ class FreqaiExampleStrategy(IStrategy):
         if "close" not in dataframe.columns:
             logger.error("数据框缺少必要的 'close' 列")
             raise ValueError("数据框缺少必要的 'close' 列")
-        
+
         try:
             label_period = self.freqai_info["feature_parameters"]["label_period_candles"]
-            # 生成 %-volatility 特征
-            dataframe["%-volatility"] = dataframe["close"].pct_change().rolling(20).std()
-            
-            # 单一回归目标
-            dataframe["&-buy_rsi"] = ta.RSI(dataframe, timeperiod=14).shift(-label_period)
-            
-            # 数据清理
+            if "%-volatility" not in dataframe.columns:
+                logger.warning("缺少 %-volatility 列，强制重新生成")
+                dataframe["%-volatility"] = dataframe["close"].pct_change().rolling(20, min_periods=1).std()
+                dataframe["%-volatility"] = dataframe["%-volatility"].replace([np.inf, -np.inf], 0)
+                dataframe["%-volatility"] = dataframe["%-volatility"].ffill()
+                dataframe["%-volatility"] = dataframe["%-volatility"].fillna(0)
+
+# 移除 shift(-label_period)，改为使用当前及过去的数据
+            dataframe["&-buy_rsi"] = ta.RSI(dataframe, timeperiod=14)
+            dataframe["&-buy_rsi"] = dataframe["&-buy_rsi"].rolling(window=label_period).mean().ffill().bfill()
+
             for col in ["&-buy_rsi", "%-volatility"]:
-                dataframe[col].replace([np.inf, -np.inf], 0, inplace=True)
-                dataframe[col].fillna(method='ffill', inplace=True)
-                dataframe[col].fillna(0, inplace=True)
+                dataframe[col] = dataframe[col].replace([np.inf, -np.inf], 0)
+                dataframe[col] = dataframe[col].ffill()
+                dataframe[col] = dataframe[col].fillna(0)
                 if dataframe[col].isna().any():
-                    logger.warning(f"目标列 {col} 仍包含 NaN，检查数据生成逻辑")
+                    logger.warning(f"目标列 {col} 仍包含 NaN，数据预览：\n{dataframe[col].tail(10)}")
         except Exception as e:
             logger.error(f"创建 FreqAI 目标失败：{str(e)}")
             raise
-        
+
         logger.info(f"目标列预览：\n{dataframe[['&-buy_rsi']].head().to_string()}")
         return dataframe
 
     def populate_indicators(self, dataframe: DataFrame, metadata: dict) -> DataFrame:
         logger.info(f"处理交易对：{metadata['pair']}")
+        logger.debug(f"输入特征列：{list(dataframe.columns)}")
         dataframe = self.freqai.start(dataframe, metadata, self)
-        
-        # 计算传统指标
+        logger.debug(f"FreqAI 输出特征列：{list(dataframe.columns)}")
+
         dataframe["rsi"] = ta.RSI(dataframe, timeperiod=14)
         bollinger = qtpylib.bollinger_bands(qtpylib.typical_price(dataframe), window=20, stds=2)
         dataframe["bb_lowerband"] = bollinger["lower"]
         dataframe["bb_middleband"] = bollinger["mid"]
         dataframe["bb_upperband"] = bollinger["upper"]
         dataframe["tema"] = ta.TEMA(dataframe, timeperiod=9)
-        
-        # 生成 up_or_down 信号（非 FreqAI 目标）
+
         label_period = self.freqai_info["feature_parameters"]["label_period_candles"]
+# 使用滚动窗口而非未来函数来生成 up_or_down 列
         dataframe["up_or_down"] = np.where(
-            dataframe["close"].shift(-label_period) > dataframe["close"], 1, 0
+            dataframe["close"].rolling(window=label_period).mean() > dataframe["close"], 1, 0
         )
-        
-        # 动态设置参数
+
         if "&-buy_rsi" in dataframe.columns:
-            # 派生其他目标
+            if "%-volatility" not in dataframe.columns:
+                logger.warning("缺少 %-volatility 列，强制重新生成")
+                dataframe["%-volatility"] = dataframe["close"].pct_change().rolling(20, min_periods=1).std()
+                dataframe["%-volatility"] = dataframe["%-volatility"].replace([np.inf, -np.inf], 0)
+                dataframe["%-volatility"] = dataframe["%-volatility"].ffill()
+                dataframe["%-volatility"] = dataframe["%-volatility"].fillna(0)
+
             dataframe["&-sell_rsi"] = dataframe["&-buy_rsi"] + 30
-            dataframe["%-volatility"] = dataframe["close"].pct_change().rolling(20).std()
-            dataframe["&-stoploss"] = -0.1 - (dataframe["%-volatility"] * 10).clip(0, 0.25)
-            dataframe["&-roi_0"] = (dataframe["close"].shift(-label_period) / dataframe["close"] - 1).clip(0, 0.2)
-            
-            # 限制预测值，添加平滑
+            dataframe["&-stoploss"] = self.stoploss - (dataframe["%-volatility"] * 5).clip(-0.05, 0.05)
+            dataframe["&-roi_0"] = (dataframe["close"].rolling(window=label_period).mean() / dataframe["close"] - 1).clip(0, 0.2)
+
+            for col in ["&-buy_rsi", "&-sell_rsi", "&-stoploss", "&-roi_0"]:
+                dataframe[col] = dataframe[col].replace([np.inf, -np.inf], 0)
+                dataframe[col] = dataframe[col].ffill()
+                dataframe[col] = dataframe[col].fillna(0)
+
             dataframe["buy_rsi_pred"] = dataframe["&-buy_rsi"].rolling(5).mean().clip(10, 50)
-            dataframe["buy_rsi_pred"].fillna(dataframe["buy_rsi_pred"].mean(), inplace=True)
-            if dataframe["buy_rsi_pred"].isna().any():
-                logger.warning("buy_rsi_pred 列包含 NaN，已填充为默认值")
             dataframe["sell_rsi_pred"] = dataframe["&-sell_rsi"].rolling(5).mean().clip(50, 90)
-            dataframe["sell_rsi_pred"].fillna(dataframe["sell_rsi_pred"].mean(), inplace=True)
-            if dataframe["sell_rsi_pred"].isna().any():
-                logger.warning("sell_rsi_pred 列包含 NaN，已填充为默认值")
             dataframe["stoploss_pred"] = dataframe["&-stoploss"].clip(-0.35, -0.1)
-            dataframe["stoploss_pred"].fillna(dataframe["stoploss_pred"].mean(), inplace=True)
-            if dataframe["stoploss_pred"].isna().any():
-                logger.warning("stoploss_pred 列包含 NaN，已填充为默认值")
             dataframe["roi_0_pred"] = dataframe["&-roi_0"].clip(0.01, 0.2)
-            dataframe["roi_0_pred"].fillna(dataframe["roi_0_pred"].mean(), inplace=True)
-            if dataframe["roi_0_pred"].isna().any():
-                logger.warning("roi_0_pred 列包含 NaN，已填充为默认值")
-                        
-            # 检查预测值
-            for col in ["buy_rsi_pred", "sell_rsi_pred", "stoploss_pred", "roi_0_pred", "&-sell_rsi", "&-stoploss", "&-roi_0"]:
+
+            for col in ["buy_rsi_pred", "sell_rsi_pred", "stoploss_pred", "roi_0_pred"]:
                 if dataframe[col].isna().any():
                     logger.warning(f"列 {col} 包含 NaN，填充为默认值")
-                    dataframe[col].fillna(dataframe[col].mean(), inplace=True)
-            
-            # 动态追踪止盈
+                    dataframe[col] = dataframe[col].ffill()
+                    dataframe[col] = dataframe[col].fillna(dataframe[col].mean())
+
             dataframe["trailing_stop_positive"] = (dataframe["roi_0_pred"] * 0.5).clip(0.01, 0.3)
             dataframe["trailing_stop_positive_offset"] = (dataframe["roi_0_pred"] * 0.75).clip(0.02, 0.4)
-            
-            # 设置策略级参数
+
             self.buy_rsi.value = float(dataframe["buy_rsi_pred"].iloc[-1])
             self.sell_rsi.value = float(dataframe["sell_rsi_pred"].iloc[-1])
             self.stoploss = float(self.stoploss_param.value)
@@ -192,19 +200,22 @@ class FreqaiExampleStrategy(IStrategy):
                 30: float(self.roi_30.value),
                 60: 0
             }
-            self.trailing_stop_positive = float(dataframe["trailing_stop_positive"].iloc[-1])
-            self.trailing_stop_positive_offset = float(dataframe["trailing_stop_positive_offset"].iloc[-1])
-            
+            self.trailing_stop_positive = float(self.trailing_stop_positive_param.value)
+            self.trailing_stop_positive_offset = float(self.trailing_stop_positive_offset_param.value)
+
             logger.info(f"动态参数：buy_rsi={self.buy_rsi.value}, sell_rsi={self.sell_rsi.value}, "
                         f"stoploss={self.stoploss}, trailing_stop_positive={self.trailing_stop_positive}")
-        
-        dataframe.replace([np.inf, -np.inf], 0, inplace=True)
-        dataframe.fillna(method='ffill', inplace=True)
-        dataframe.fillna(0, inplace=True)
-        
+        else:
+            logger.warning(f"&-buy_rsi 列缺失，跳过 FreqAI 预测逻辑，检查 freqai.start 输出")
+
+        dataframe = dataframe.replace([np.inf, -np.inf], 0)
+        dataframe = dataframe.ffill()
+        dataframe = dataframe.fillna(0)
+
         logger.info(f"up_or_down 值统计：\n{dataframe['up_or_down'].value_counts().to_string()}")
         logger.info(f"do_predict 值统计：\n{dataframe['do_predict'].value_counts().to_string()}")
-        
+        logger.debug(f"最终特征列：{list(dataframe.columns)}")
+
         return dataframe
 
     def populate_entry_trend(self, df: DataFrame, metadata: dict) -> DataFrame:
@@ -241,9 +252,28 @@ class FreqaiExampleStrategy(IStrategy):
         self, pair: str, order_type: str, amount: float, rate: float,
         time_in_force: str, current_time, entry_tag, side: str, **kwargs
     ) -> bool:
-        df, _ = self.dp.get_analyzed_dataframe(pair, self.timeframe)
-        last_candle = df.iloc[-1].squeeze()
-        if side == "long":
-            if rate > (last_candle["close"] * (1 + 0.0025)):
+        try:
+            df, _ = self.dp.get_analyzed_dataframe(pair, self.timeframe)
+            if df is None or df.empty:
+                logger.warning(f"无法获取 {pair} 的分析数据，拒绝交易")
                 return False
-        return True
+
+            last_candle = df.iloc[-1].squeeze()
+            if "close" not in last_candle or np.isnan(last_candle["close"]):
+                logger.warning(f"{pair} 的最新 K 线缺少有效 close 价格，拒绝交易")
+                return False
+
+            if side == "long":
+                max_rate = last_candle["close"] * (1 + 0.0025)  # 0.25% 滑点阈值
+                if rate > max_rate:
+                    logger.debug(f"拒绝 {pair} 的买入，价格 {rate} 超过最大允许价格 {max_rate}")
+                    return False
+            elif side == "short":
+                logger.warning(f"{pair} 尝试做空，但策略不支持做空 (can_short={self.can_short})")
+                return False
+
+            logger.debug(f"确认 {pair} 的交易：side={side}, rate={rate}, close={last_candle['close']}")
+            return True
+        except Exception as e:
+            logger.error(f"确认 {pair} 交易时出错：{str(e)}")
+            return False

freqtrade/templates/sample_hyperopt_loss.py

@@ -1,57 +0,0 @@
-from datetime import datetime
-from math import exp
-
-from pandas import DataFrame
-
-from freqtrade.constants import Config
-from freqtrade.optimize.hyperopt import IHyperOptLoss
-
-
-# Define some constants:
-
-# set TARGET_TRADES to suit your number concurrent trades so its realistic
-# to the number of days
-TARGET_TRADES = 600
-# This is assumed to be expected avg profit * expected trade count.
-# For example, for 0.35% avg per trade (or 0.0035 as ratio) and 1100 trades,
-# self.expected_max_profit = 3.85
-# Check that the reported Σ% values do not exceed this!
-# Note, this is ratio. 3.85 stated above means 385Σ%.
-EXPECTED_MAX_PROFIT = 3.0
-
-# max average trade duration in minutes
-# if eval ends with higher value, we consider it a failed eval
-MAX_ACCEPTED_TRADE_DURATION = 300
-
-
-class SampleHyperOptLoss(IHyperOptLoss):
-    """
-    Defines the default loss function for hyperopt
-    This is intended to give you some inspiration for your own loss function.
-
-    The Function needs to return a number (float) - which becomes smaller for better backtest
-    results.
-    """
-
-    @staticmethod
-    def hyperopt_loss_function(
-        results: DataFrame,
-        trade_count: int,
-        min_date: datetime,
-        max_date: datetime,
-        config: Config,
-        processed: dict[str, DataFrame],
-        *args,
-        **kwargs,
-    ) -> float:
-        """
-        Objective function, returns smaller number for better results
-        """
-        total_profit = results["profit_ratio"].sum()
-        trade_duration = results["trade_duration"].mean()
-
-        trade_loss = 1 - 0.25 * exp(-((trade_count - TARGET_TRADES) ** 2) / 10**5.8)
-        profit_loss = max(0, 1 - total_profit / EXPECTED_MAX_PROFIT)
-        duration_loss = 0.4 * min(trade_duration / MAX_ACCEPTED_TRADE_DURATION, 1)
-        result = trade_loss + profit_loss + duration_loss
-        return result

freqtrade/templates/sample_strategy.py

@@ -1,426 +0,0 @@
-# pragma pylint: disable=missing-docstring, invalid-name, pointless-string-statement
-# flake8: noqa: F401
-# isort: skip_file
-# --- Do not remove these imports ---
-import numpy as np
-import pandas as pd
-from datetime import datetime, timedelta, timezone
-from pandas import DataFrame
-from typing import Optional, Union
-
-from freqtrade.strategy import (
-    IStrategy,
-    Trade,
-    Order,
-    PairLocks,
-    informative,  # @informative decorator
-    # Hyperopt Parameters
-    BooleanParameter,
-    CategoricalParameter,
-    DecimalParameter,
-    IntParameter,
-    RealParameter,
-    # timeframe helpers
-    timeframe_to_minutes,
-    timeframe_to_next_date,
-    timeframe_to_prev_date,
-    # Strategy helper functions
-    merge_informative_pair,
-    stoploss_from_absolute,
-    stoploss_from_open,
-)
-
-# --------------------------------
-# Add your lib to import here
-import talib.abstract as ta
-from technical import qtpylib
-
-
-# This class is a sample. Feel free to customize it.
-class SampleStrategy(IStrategy):
-    """
-    This is a sample strategy to inspire you.
-    More information in https://www.freqtrade.io/en/latest/strategy-customization/
-
-    You can:
-        :return: a Dataframe with all mandatory indicators for the strategies
-    - Rename the class name (Do not forget to update class_name)
-    - Add any methods you want to build your strategy
-    - Add any lib you need to build your strategy
-
-    You must keep:
-    - the lib in the section "Do not remove these libs"
-    - the methods: populate_indicators, populate_entry_trend, populate_exit_trend
-    You should keep:
-    - timeframe, minimal_roi, stoploss, trailing_*
-    """
-
-    # Strategy interface version - allow new iterations of the strategy interface.
-    # Check the documentation or the Sample strategy to get the latest version.
-    INTERFACE_VERSION = 3
-
-    # Can this strategy go short?
-    can_short: bool = False
-
-    # Minimal ROI designed for the strategy.
-    # This attribute will be overridden if the config file contains "minimal_roi".
-    minimal_roi = {
-        # "120": 0.0,  # exit after 120 minutes at break even
-        "60": 0.01,
-        "30": 0.02,
-        "0": 0.04,
-    }
-
-    # Optimal stoploss designed for the strategy.
-    # This attribute will be overridden if the config file contains "stoploss".
-    stoploss = -0.10
-
-    # Trailing stoploss
-    trailing_stop = False
-    # trailing_only_offset_is_reached = False
-    # trailing_stop_positive = 0.01
-    # trailing_stop_positive_offset = 0.0  # Disabled / not configured
-
-    # Optimal timeframe for the strategy.
-    timeframe = "5m"
-
-    # Run "populate_indicators()" only for new candle.
-    process_only_new_candles = True
-
-    # These values can be overridden in the config.
-    use_exit_signal = True
-    exit_profit_only = False
-    ignore_roi_if_entry_signal = False
-
-    # Hyperoptable parameters
-    buy_rsi = IntParameter(low=1, high=50, default=30, space="buy", optimize=True, load=True)
-    sell_rsi = IntParameter(low=50, high=100, default=70, space="sell", optimize=True, load=True)
-    short_rsi = IntParameter(low=51, high=100, default=70, space="sell", optimize=True, load=True)
-    exit_short_rsi = IntParameter(low=1, high=50, default=30, space="buy", optimize=True, load=True)
-
-    # Number of candles the strategy requires before producing valid signals
-    startup_candle_count: int = 200
-
-    # Optional order type mapping.
-    order_types = {
-        "entry": "limit",
-        "exit": "limit",
-        "stoploss": "market",
-        "stoploss_on_exchange": False,
-    }
-
-    # Optional order time in force.
-    order_time_in_force = {"entry": "GTC", "exit": "GTC"}
-
-    plot_config = {
-        "main_plot": {
-            "tema": {},
-            "sar": {"color": "white"},
-        },
-        "subplots": {
-            "MACD": {
-                "macd": {"color": "blue"},
-                "macdsignal": {"color": "orange"},
-            },
-            "RSI": {
-                "rsi": {"color": "red"},
-            },
-        },
-    }
-
-    def informative_pairs(self):
-        """
-        Define additional, informative pair/interval combinations to be cached from the exchange.
-        These pair/interval combinations are non-tradeable, unless they are part
-        of the whitelist as well.
-        For more information, please consult the documentation
-        :return: List of tuples in the format (pair, interval)
-            Sample: return [("ETH/USDT", "5m"),
-                            ("BTC/USDT", "15m"),
-                            ]
-        """
-        return []
-
-    def populate_indicators(self, dataframe: DataFrame, metadata: dict) -> DataFrame:
-        """
-        Adds several different TA indicators to the given DataFrame
-
-        Performance Note: For the best performance be frugal on the number of indicators
-        you are using. Let uncomment only the indicator you are using in your strategies
-        or your hyperopt configuration, otherwise you will waste your memory and CPU usage.
-        :param dataframe: Dataframe with data from the exchange
-        :param metadata: Additional information, like the currently traded pair
-        :return: a Dataframe with all mandatory indicators for the strategies
-        """
-
-        # Momentum Indicators
-        # ------------------------------------
-
-        # ADX
-        dataframe["adx"] = ta.ADX(dataframe)
-
-        # # Plus Directional Indicator / Movement
-        # dataframe['plus_dm'] = ta.PLUS_DM(dataframe)
-        # dataframe['plus_di'] = ta.PLUS_DI(dataframe)
-
-        # # Minus Directional Indicator / Movement
-        # dataframe['minus_dm'] = ta.MINUS_DM(dataframe)
-        # dataframe['minus_di'] = ta.MINUS_DI(dataframe)
-
-        # # Aroon, Aroon Oscillator
-        # aroon = ta.AROON(dataframe)
-        # dataframe['aroonup'] = aroon['aroonup']
-        # dataframe['aroondown'] = aroon['aroondown']
-        # dataframe['aroonosc'] = ta.AROONOSC(dataframe)
-
-        # # Awesome Oscillator
-        # dataframe['ao'] = qtpylib.awesome_oscillator(dataframe)
-
-        # # Keltner Channel
-        # keltner = qtpylib.keltner_channel(dataframe)
-        # dataframe["kc_upperband"] = keltner["upper"]
-        # dataframe["kc_lowerband"] = keltner["lower"]
-        # dataframe["kc_middleband"] = keltner["mid"]
-        # dataframe["kc_percent"] = (
-        #     (dataframe["close"] - dataframe["kc_lowerband"]) /
-        #     (dataframe["kc_upperband"] - dataframe["kc_lowerband"])
-        # )
-        # dataframe["kc_width"] = (
-        #     (dataframe["kc_upperband"] - dataframe["kc_lowerband"]) / dataframe["kc_middleband"]
-        # )
-
-        # # Ultimate Oscillator
-        # dataframe['uo'] = ta.ULTOSC(dataframe)
-
-        # # Commodity Channel Index: values [Oversold:-100, Overbought:100]
-        # dataframe['cci'] = ta.CCI(dataframe)
-
-        # RSI
-        dataframe["rsi"] = ta.RSI(dataframe)
-
-        # # Inverse Fisher transform on RSI: values [-1.0, 1.0] (https://goo.gl/2JGGoy)
-        # rsi = 0.1 * (dataframe['rsi'] - 50)
-        # dataframe['fisher_rsi'] = (np.exp(2 * rsi) - 1) / (np.exp(2 * rsi) + 1)
-
-        # # Inverse Fisher transform on RSI normalized: values [0.0, 100.0] (https://goo.gl/2JGGoy)
-        # dataframe['fisher_rsi_norma'] = 50 * (dataframe['fisher_rsi'] + 1)
-
-        # # Stochastic Slow
-        # stoch = ta.STOCH(dataframe)
-        # dataframe['slowd'] = stoch['slowd']
-        # dataframe['slowk'] = stoch['slowk']
-
-        # Stochastic Fast
-        stoch_fast = ta.STOCHF(dataframe)
-        dataframe["fastd"] = stoch_fast["fastd"]
-        dataframe["fastk"] = stoch_fast["fastk"]
-
-        # # Stochastic RSI
-        # Please read https://github.com/freqtrade/freqtrade/issues/2961 before using this.
-        # STOCHRSI is NOT aligned with tradingview, which may result in non-expected results.
-        # stoch_rsi = ta.STOCHRSI(dataframe)
-        # dataframe['fastd_rsi'] = stoch_rsi['fastd']
-        # dataframe['fastk_rsi'] = stoch_rsi['fastk']
-
-        # MACD
-        macd = ta.MACD(dataframe)
-        dataframe["macd"] = macd["macd"]
-        dataframe["macdsignal"] = macd["macdsignal"]
-        dataframe["macdhist"] = macd["macdhist"]
-
-        # MFI
-        dataframe["mfi"] = ta.MFI(dataframe)
-
-        # # ROC
-        # dataframe['roc'] = ta.ROC(dataframe)
-
-        # Overlap Studies
-        # ------------------------------------
-
-        # Bollinger Bands
-        bollinger = qtpylib.bollinger_bands(qtpylib.typical_price(dataframe), window=20, stds=2)
-        dataframe["bb_lowerband"] = bollinger["lower"]
-        dataframe["bb_middleband"] = bollinger["mid"]
-        dataframe["bb_upperband"] = bollinger["upper"]
-        dataframe["bb_percent"] = (dataframe["close"] - dataframe["bb_lowerband"]) / (
-            dataframe["bb_upperband"] - dataframe["bb_lowerband"]
-        )
-        dataframe["bb_width"] = (dataframe["bb_upperband"] - dataframe["bb_lowerband"]) / dataframe[
-            "bb_middleband"
-        ]
-
-        # Bollinger Bands - Weighted (EMA based instead of SMA)
-        # weighted_bollinger = qtpylib.weighted_bollinger_bands(
-        #     qtpylib.typical_price(dataframe), window=20, stds=2
-        # )
-        # dataframe["wbb_upperband"] = weighted_bollinger["upper"]
-        # dataframe["wbb_lowerband"] = weighted_bollinger["lower"]
-        # dataframe["wbb_middleband"] = weighted_bollinger["mid"]
-        # dataframe["wbb_percent"] = (
-        #     (dataframe["close"] - dataframe["wbb_lowerband"]) /
-        #     (dataframe["wbb_upperband"] - dataframe["wbb_lowerband"])
-        # )
-        # dataframe["wbb_width"] = (
-        #     (dataframe["wbb_upperband"] - dataframe["wbb_lowerband"]) /
-        #     dataframe["wbb_middleband"]
-        # )
-
-        # # EMA - Exponential Moving Average
-        # dataframe['ema3'] = ta.EMA(dataframe, timeperiod=3)
-        # dataframe['ema5'] = ta.EMA(dataframe, timeperiod=5)
-        # dataframe['ema10'] = ta.EMA(dataframe, timeperiod=10)
-        # dataframe['ema21'] = ta.EMA(dataframe, timeperiod=21)
-        # dataframe['ema50'] = ta.EMA(dataframe, timeperiod=50)
-        # dataframe['ema100'] = ta.EMA(dataframe, timeperiod=100)
-
-        # # SMA - Simple Moving Average
-        # dataframe['sma3'] = ta.SMA(dataframe, timeperiod=3)
-        # dataframe['sma5'] = ta.SMA(dataframe, timeperiod=5)
-        # dataframe['sma10'] = ta.SMA(dataframe, timeperiod=10)
-        # dataframe['sma21'] = ta.SMA(dataframe, timeperiod=21)
-        # dataframe['sma50'] = ta.SMA(dataframe, timeperiod=50)
-        # dataframe['sma100'] = ta.SMA(dataframe, timeperiod=100)
-
-        # Parabolic SAR
-        dataframe["sar"] = ta.SAR(dataframe)
-
-        # TEMA - Triple Exponential Moving Average
-        dataframe["tema"] = ta.TEMA(dataframe, timeperiod=9)
-
-        # Cycle Indicator
-        # ------------------------------------
-        # Hilbert Transform Indicator - SineWave
-        hilbert = ta.HT_SINE(dataframe)
-        dataframe["htsine"] = hilbert["sine"]
-        dataframe["htleadsine"] = hilbert["leadsine"]
-
-        # Pattern Recognition - Bullish candlestick patterns
-        # ------------------------------------
-        # # Hammer: values [0, 100]
-        # dataframe['CDLHAMMER'] = ta.CDLHAMMER(dataframe)
-        # # Inverted Hammer: values [0, 100]
-        # dataframe['CDLINVERTEDHAMMER'] = ta.CDLINVERTEDHAMMER(dataframe)
-        # # Dragonfly Doji: values [0, 100]
-        # dataframe['CDLDRAGONFLYDOJI'] = ta.CDLDRAGONFLYDOJI(dataframe)
-        # # Piercing Line: values [0, 100]
-        # dataframe['CDLPIERCING'] = ta.CDLPIERCING(dataframe) # values [0, 100]
-        # # Morningstar: values [0, 100]
-        # dataframe['CDLMORNINGSTAR'] = ta.CDLMORNINGSTAR(dataframe) # values [0, 100]
-        # # Three White Soldiers: values [0, 100]
-        # dataframe['CDL3WHITESOLDIERS'] = ta.CDL3WHITESOLDIERS(dataframe) # values [0, 100]
-
-        # Pattern Recognition - Bearish candlestick patterns
-        # ------------------------------------
-        # # Hanging Man: values [0, 100]
-        # dataframe['CDLHANGINGMAN'] = ta.CDLHANGINGMAN(dataframe)
-        # # Shooting Star: values [0, 100]
-        # dataframe['CDLSHOOTINGSTAR'] = ta.CDLSHOOTINGSTAR(dataframe)
-        # # Gravestone Doji: values [0, 100]
-        # dataframe['CDLGRAVESTONEDOJI'] = ta.CDLGRAVESTONEDOJI(dataframe)
-        # # Dark Cloud Cover: values [0, 100]
-        # dataframe['CDLDARKCLOUDCOVER'] = ta.CDLDARKCLOUDCOVER(dataframe)
-        # # Evening Doji Star: values [0, 100]
-        # dataframe['CDLEVENINGDOJISTAR'] = ta.CDLEVENINGDOJISTAR(dataframe)
-        # # Evening Star: values [0, 100]
-        # dataframe['CDLEVENINGSTAR'] = ta.CDLEVENINGSTAR(dataframe)
-
-        # Pattern Recognition - Bullish/Bearish candlestick patterns
-        # ------------------------------------
-        # # Three Line Strike: values [0, -100, 100]
-        # dataframe['CDL3LINESTRIKE'] = ta.CDL3LINESTRIKE(dataframe)
-        # # Spinning Top: values [0, -100, 100]
-        # dataframe['CDLSPINNINGTOP'] = ta.CDLSPINNINGTOP(dataframe) # values [0, -100, 100]
-        # # Engulfing: values [0, -100, 100]
-        # dataframe['CDLENGULFING'] = ta.CDLENGULFING(dataframe) # values [0, -100, 100]
-        # # Harami: values [0, -100, 100]
-        # dataframe['CDLHARAMI'] = ta.CDLHARAMI(dataframe) # values [0, -100, 100]
-        # # Three Outside Up/Down: values [0, -100, 100]
-        # dataframe['CDL3OUTSIDE'] = ta.CDL3OUTSIDE(dataframe) # values [0, -100, 100]
-        # # Three Inside Up/Down: values [0, -100, 100]
-        # dataframe['CDL3INSIDE'] = ta.CDL3INSIDE(dataframe) # values [0, -100, 100]
-
-        # # Chart type
-        # # ------------------------------------
-        # # Heikin Ashi Strategy
-        # heikinashi = qtpylib.heikinashi(dataframe)
-        # dataframe['ha_open'] = heikinashi['open']
-        # dataframe['ha_close'] = heikinashi['close']
-        # dataframe['ha_high'] = heikinashi['high']
-        # dataframe['ha_low'] = heikinashi['low']
-
-        # Retrieve best bid and best ask from the orderbook
-        # ------------------------------------
-        """
-        # first check if dataprovider is available
-        if self.dp:
-            if self.dp.runmode.value in ('live', 'dry_run'):
-                ob = self.dp.orderbook(metadata['pair'], 1)
-                dataframe['best_bid'] = ob['bids'][0][0]
-                dataframe['best_ask'] = ob['asks'][0][0]
-        """
-
-        return dataframe
-
-    def populate_entry_trend(self, dataframe: DataFrame, metadata: dict) -> DataFrame:
-        """
-        Based on TA indicators, populates the entry signal for the given dataframe
-        :param dataframe: DataFrame
-        :param metadata: Additional information, like the currently traded pair
-        :return: DataFrame with entry columns populated
-        """
-        dataframe.loc[
-            (
-                # Signal: RSI crosses above 30
-                (qtpylib.crossed_above(dataframe["rsi"], self.buy_rsi.value))
-                & (dataframe["tema"] <= dataframe["bb_middleband"])  # Guard: tema below BB middle
-                & (dataframe["tema"] > dataframe["tema"].shift(1))  # Guard: tema is raising
-                & (dataframe["volume"] > 0)  # Make sure Volume is not 0
-            ),
-            "enter_long",
-        ] = 1
-
-        dataframe.loc[
-            (
-                # Signal: RSI crosses above 70
-                (qtpylib.crossed_above(dataframe["rsi"], self.short_rsi.value))
-                & (dataframe["tema"] > dataframe["bb_middleband"])  # Guard: tema above BB middle
-                & (dataframe["tema"] < dataframe["tema"].shift(1))  # Guard: tema is falling
-                & (dataframe["volume"] > 0)  # Make sure Volume is not 0
-            ),
-            "enter_short",
-        ] = 1
-
-        return dataframe
-
-    def populate_exit_trend(self, dataframe: DataFrame, metadata: dict) -> DataFrame:
-        """
-        Based on TA indicators, populates the exit signal for the given dataframe
-        :param dataframe: DataFrame
-        :param metadata: Additional information, like the currently traded pair
-        :return: DataFrame with exit columns populated
-        """
-        dataframe.loc[
-            (
-                # Signal: RSI crosses above 70
-                (qtpylib.crossed_above(dataframe["rsi"], self.sell_rsi.value))
-                & (dataframe["tema"] > dataframe["bb_middleband"])  # Guard: tema above BB middle
-                & (dataframe["tema"] < dataframe["tema"].shift(1))  # Guard: tema is falling
-                & (dataframe["volume"] > 0)  # Make sure Volume is not 0
-            ),
-            "exit_long",
-        ] = 1
-
-        dataframe.loc[
-            (
-                # Signal: RSI crosses above 30
-                (qtpylib.crossed_above(dataframe["rsi"], self.exit_short_rsi.value))
-                &
-                # Guard: tema below BB middle
-                (dataframe["tema"] <= dataframe["bb_middleband"])
-                & (dataframe["tema"] > dataframe["tema"].shift(1))  # Guard: tema is raising
-                & (dataframe["volume"] > 0)  # Make sure Volume is not 0
-            ),
-            "exit_short",
-        ] = 1
-
-        return dataframe

freqtrade/templates/strategy_analysis_example.ipynb

@@ -1,480 +0,0 @@
-{
- "cells": [
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "# Strategy analysis example\n",
-    "\n",
-    "Debugging a strategy can be time-consuming. Freqtrade offers helper functions to visualize raw data.\n",
-    "The following assumes you work with SampleStrategy, data for 5m timeframe from Binance and have downloaded them into the data directory in the default location.\n",
-    "Please follow the [documentation](https://www.freqtrade.io/en/stable/data-download/) for more details."
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "## Setup\n",
-    "\n",
-    "### Change Working directory to repository root"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "import os\n",
-    "from pathlib import Path\n",
-    "\n",
-    "\n",
-    "# Change directory\n",
-    "# Modify this cell to insure that the output shows the correct path.\n",
-    "# Define all paths relative to the project root shown in the cell output\n",
-    "project_root = \"somedir/freqtrade\"\n",
-    "i = 0\n",
-    "try:\n",
-    "    os.chdir(project_root)\n",
-    "    if not Path(\"LICENSE\").is_file():\n",
-    "        i = 0\n",
-    "        while i < 4 and (not Path(\"LICENSE\").is_file()):\n",
-    "            os.chdir(Path(Path.cwd(), \"../\"))\n",
-    "            i += 1\n",
-    "        project_root = Path.cwd()\n",
-    "except FileNotFoundError:\n",
-    "    print(\"Please define the project root relative to the current directory\")\n",
-    "print(Path.cwd())"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "### Configure Freqtrade environment"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "from freqtrade.configuration import Configuration\n",
-    "\n",
-    "\n",
-    "# Customize these according to your needs.\n",
-    "\n",
-    "# Initialize empty configuration object\n",
-    "config = Configuration.from_files([])\n",
-    "# Optionally (recommended), use existing configuration file\n",
-    "# config = Configuration.from_files([\"user_data/config.json\"])\n",
-    "\n",
-    "# Define some constants\n",
-    "config[\"timeframe\"] = \"5m\"\n",
-    "# Name of the strategy class\n",
-    "config[\"strategy\"] = \"SampleStrategy\"\n",
-    "# Location of the data\n",
-    "data_location = config[\"datadir\"]\n",
-    "# Pair to analyze - Only use one pair here\n",
-    "pair = \"BTC/USDT\""
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "# Load data using values set above\n",
-    "from freqtrade.data.history import load_pair_history\n",
-    "from freqtrade.enums import CandleType\n",
-    "\n",
-    "\n",
-    "candles = load_pair_history(\n",
-    "    datadir=data_location,\n",
-    "    timeframe=config[\"timeframe\"],\n",
-    "    pair=pair,\n",
-    "    data_format=\"json\",  # Make sure to update this to your data\n",
-    "    candle_type=CandleType.SPOT,\n",
-    ")\n",
-    "\n",
-    "# Confirm success\n",
-    "print(f\"Loaded {len(candles)} rows of data for {pair} from {data_location}\")\n",
-    "candles.head()"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "## Load and run strategy\n",
-    "* Rerun each time the strategy file is changed"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "# Load strategy using values set above\n",
-    "from freqtrade.data.dataprovider import DataProvider\n",
-    "from freqtrade.resolvers import StrategyResolver\n",
-    "\n",
-    "\n",
-    "strategy = StrategyResolver.load_strategy(config)\n",
-    "strategy.dp = DataProvider(config, None, None)\n",
-    "strategy.ft_bot_start()\n",
-    "\n",
-    "# Generate buy/sell signals using strategy\n",
-    "df = strategy.analyze_ticker(candles, {\"pair\": pair})\n",
-    "df.tail()"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "### Display the trade details\n",
-    "\n",
-    "* Note that using `data.head()` would also work, however most indicators have some \"startup\" data at the top of the dataframe.\n",
-    "* Some possible problems\n",
-    "    * Columns with NaN values at the end of the dataframe\n",
-    "    * Columns used in `crossed*()` functions with completely different units\n",
-    "* Comparison with full backtest\n",
-    "    * having 200 buy signals as output for one pair from `analyze_ticker()` does not necessarily mean that 200 trades will be made during backtesting.\n",
-    "    * Assuming you use only one condition such as, `df['rsi'] < 30` as buy condition, this will generate multiple \"buy\" signals for each pair in sequence (until rsi returns > 29). The bot will only buy on the first of these signals (and also only if a trade-slot (\"max_open_trades\") is still available), or on one of the middle signals, as soon as a \"slot\" becomes available.  \n"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "# Report results\n",
-    "print(f\"Generated {df['enter_long'].sum()} entry signals\")\n",
-    "data = df.set_index(\"date\", drop=False)\n",
-    "data.tail()"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "## Load existing objects into a Jupyter notebook\n",
-    "\n",
-    "The following cells assume that you have already generated data using the cli.  \n",
-    "They will allow you to drill deeper into your results, and perform analysis which otherwise would make the output very difficult to digest due to information overload."
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "### Load backtest results to pandas dataframe\n",
-    "\n",
-    "Analyze a trades dataframe (also used below for plotting)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "from freqtrade.data.btanalysis import load_backtest_data, load_backtest_stats\n",
-    "\n",
-    "\n",
-    "# if backtest_dir points to a directory, it'll automatically load the last backtest file.\n",
-    "backtest_dir = config[\"user_data_dir\"] / \"backtest_results\"\n",
-    "# backtest_dir can also point to a specific file\n",
-    "# backtest_dir = (\n",
-    "#   config[\"user_data_dir\"] / \"backtest_results/backtest-result-2020-07-01_20-04-22.json\"\n",
-    "# )"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "# You can get the full backtest statistics by using the following command.\n",
-    "# This contains all information used to generate the backtest result.\n",
-    "stats = load_backtest_stats(backtest_dir)\n",
-    "\n",
-    "strategy = \"SampleStrategy\"\n",
-    "# All statistics are available per strategy, so if `--strategy-list` was used during backtest,\n",
-    "# this will be reflected here as well.\n",
-    "# Example usages:\n",
-    "print(stats[\"strategy\"][strategy][\"results_per_pair\"])\n",
-    "# Get pairlist used for this backtest\n",
-    "print(stats[\"strategy\"][strategy][\"pairlist\"])\n",
-    "# Get market change (average change of all pairs from start to end of the backtest period)\n",
-    "print(stats[\"strategy\"][strategy][\"market_change\"])\n",
-    "# Maximum drawdown ()\n",
-    "print(stats[\"strategy\"][strategy][\"max_drawdown_abs\"])\n",
-    "# Maximum drawdown start and end\n",
-    "print(stats[\"strategy\"][strategy][\"drawdown_start\"])\n",
-    "print(stats[\"strategy\"][strategy][\"drawdown_end\"])\n",
-    "\n",
-    "\n",
-    "# Get strategy comparison (only relevant if multiple strategies were compared)\n",
-    "print(stats[\"strategy_comparison\"])"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "# Load backtested trades as dataframe\n",
-    "trades = load_backtest_data(backtest_dir)\n",
-    "\n",
-    "# Show value-counts per pair\n",
-    "trades.groupby(\"pair\")[\"exit_reason\"].value_counts()"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "## Plotting daily profit / equity line"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "# Plotting equity line (starting with 0 on day 1 and adding daily profit for each backtested day)\n",
-    "\n",
-    "import pandas as pd\n",
-    "import plotly.express as px\n",
-    "\n",
-    "from freqtrade.configuration import Configuration\n",
-    "from freqtrade.data.btanalysis import load_backtest_stats\n",
-    "\n",
-    "\n",
-    "# strategy = 'SampleStrategy'\n",
-    "# config = Configuration.from_files([\"user_data/config.json\"])\n",
-    "# backtest_dir = config[\"user_data_dir\"] / \"backtest_results\"\n",
-    "\n",
-    "stats = load_backtest_stats(backtest_dir)\n",
-    "strategy_stats = stats[\"strategy\"][strategy]\n",
-    "\n",
-    "df = pd.DataFrame(columns=[\"dates\", \"equity\"], data=strategy_stats[\"daily_profit\"])\n",
-    "df[\"equity_daily\"] = df[\"equity\"].cumsum()\n",
-    "\n",
-    "fig = px.line(df, x=\"dates\", y=\"equity_daily\")\n",
-    "fig.show()"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "### Load live trading results into a pandas dataframe\n",
-    "\n",
-    "In case you did already some trading and want to analyze your performance"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "from freqtrade.data.btanalysis import load_trades_from_db\n",
-    "\n",
-    "\n",
-    "# Fetch trades from database\n",
-    "trades = load_trades_from_db(\"sqlite:///tradesv3.sqlite\")\n",
-    "\n",
-    "# Display results\n",
-    "trades.groupby(\"pair\")[\"exit_reason\"].value_counts()"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "## Analyze the loaded trades for trade parallelism\n",
-    "This can be useful to find the best `max_open_trades` parameter, when used with backtesting in conjunction with a very high `max_open_trades` setting.\n",
-    "\n",
-    "`analyze_trade_parallelism()` returns a timeseries dataframe with an \"open_trades\" column, specifying the number of open trades for each candle."
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "from freqtrade.data.btanalysis import analyze_trade_parallelism\n",
-    "\n",
-    "\n",
-    "# Analyze the above\n",
-    "parallel_trades = analyze_trade_parallelism(trades, \"5m\")\n",
-    "\n",
-    "parallel_trades.plot()"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "## Plot results\n",
-    "\n",
-    "Freqtrade offers interactive plotting capabilities based on plotly."
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "from freqtrade.plot.plotting import generate_candlestick_graph\n",
-    "\n",
-    "\n",
-    "# Limit graph period to keep plotly quick and reactive\n",
-    "\n",
-    "# Filter trades to one pair\n",
-    "trades_red = trades.loc[trades[\"pair\"] == pair]\n",
-    "\n",
-    "data_red = data[\"2019-06-01\":\"2019-06-10\"]\n",
-    "# Generate candlestick graph\n",
-    "graph = generate_candlestick_graph(\n",
-    "    pair=pair,\n",
-    "    data=data_red,\n",
-    "    trades=trades_red,\n",
-    "    indicators1=[\"sma20\", \"ema50\", \"ema55\"],\n",
-    "    indicators2=[\"rsi\", \"macd\", \"macdsignal\", \"macdhist\"],\n",
-    ")"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "# Show graph inline\n",
-    "# graph.show()\n",
-    "\n",
-    "# Render graph in a separate window\n",
-    "graph.show(renderer=\"browser\")"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "## Plot average profit per trade as distribution graph"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "import plotly.figure_factory as ff\n",
-    "\n",
-    "\n",
-    "hist_data = [trades.profit_ratio]\n",
-    "group_labels = [\"profit_ratio\"]  # name of the dataset\n",
-    "\n",
-    "fig = ff.create_distplot(hist_data, group_labels, bin_size=0.01)\n",
-    "fig.show()"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "Feel free to submit an issue or Pull Request enhancing this document if you would like to share ideas on how to best analyze the data."
-   ]
-  }
- ],
- "metadata": {
-  "file_extension": ".py",
-  "kernelspec": {
-   "display_name": "Python 3.9.7 64-bit",
-   "language": "python",
-   "name": "python3"
-  },
-  "language_info": {
-   "codemirror_mode": {
-    "name": "ipython",
-    "version": 3
-   },
-   "file_extension": ".py",
-   "mimetype": "text/x-python",
-   "name": "python",
-   "nbconvert_exporter": "python",
-   "pygments_lexer": "ipython3",
-   "version": "3.11.4"
-  },
-  "mimetype": "text/x-python",
-  "name": "python",
-  "npconvert_exporter": "python",
-  "pygments_lexer": "ipython3",
-  "toc": {
-   "base_numbering": 1,
-   "nav_menu": {},
-   "number_sections": true,
-   "sideBar": true,
-   "skip_h1_title": false,
-   "title_cell": "Table of Contents",
-   "title_sidebar": "Contents",
-   "toc_cell": false,
-   "toc_position": {},
-   "toc_section_display": true,
-   "toc_window_display": false
-  },
-  "varInspector": {
-   "cols": {
-    "lenName": 16,
-    "lenType": 16,
-    "lenVar": 40
-   },
-   "kernels_config": {
-    "python": {
-     "delete_cmd_postfix": "",
-     "delete_cmd_prefix": "del ",
-     "library": "var_list.py",
-     "varRefreshCmd": "print(var_dic_list())"
-    },
-    "r": {
-     "delete_cmd_postfix": ") ",
-     "delete_cmd_prefix": "rm(",
-     "library": "var_list.r",
-     "varRefreshCmd": "cat(var_dic_list()) "
-    }
-   },
-   "types_to_exclude": [
-    "module",
-    "function",
-    "builtin_function_or_method",
-    "instance",
-    "_Feature"
-   ],
-   "window_display": false
-  },
-  "version": 3,
-  "vscode": {
-   "interpreter": {
-    "hash": "675f32a300d6d26767470181ad0b11dd4676bcce7ed1dd2ffe2fbc370c95fc7c"
-   }
-  }
- },
- "nbformat": 4,
- "nbformat_minor": 4
-}

md/configuration.md

@@ -0,0 +1,800 @@
+# Configure the bot
+
+Freqtrade has many configurable features and possibilities.
+By default, these settings are configured via the configuration file (see below).
+
+## The Freqtrade configuration file
+
+The bot uses a set of configuration parameters during its operation that all together conform to the bot configuration. It normally reads its configuration from a file (Freqtrade configuration file).
+
+Per default, the bot loads the configuration from the `config.json` file, located in the current working directory.
+
+You can specify a different configuration file used by the bot with the `-c/--config` command-line option.
+
+If you used the [Quick start](docker_quickstart.md#docker-quick-start) method for installing
+the bot, the installation script should have already created the default configuration file (`config.json`) for you.
+
+If the default configuration file is not created we recommend to use `freqtrade new-config --config user_data/config.json` to generate a basic configuration file.
+
+The Freqtrade configuration file is to be written in JSON format.
+
+Additionally to the standard JSON syntax, you may use one-line `// ...` and multi-line `/* ... */` comments in your configuration files and trailing commas in the lists of parameters.
+
+Do not worry if you are not familiar with JSON format -- simply open the configuration file with an editor of your choice, make some changes to the parameters you need, save your changes and, finally, restart the bot or, if it was previously stopped, run it again with the changes you made to the configuration. The bot validates the syntax of the configuration file at startup and will warn you if you made any errors editing it, pointing out problematic lines.
+
+### Environment variables
+
+Set options in the Freqtrade configuration via environment variables.
+This takes priority over the corresponding value in configuration or strategy.
+
+Environment variables must be prefixed with  `FREQTRADE__` to be loaded to the freqtrade configuration.
+
+`__` serves as level separator, so the format used should correspond to `FREQTRADE__{section}__{key}`.
+As such - an environment variable defined as  `export FREQTRADE__STAKE_AMOUNT=200` would result in `{stake_amount: 200}`.
+
+A more complex example might be `export FREQTRADE__EXCHANGE__KEY=<yourExchangeKey>` to keep your exchange key secret. This will move the value to the `exchange.key` section of the configuration.
+Using this scheme, all configuration settings will also be available as environment variables.
+
+Please note that Environment variables will overwrite corresponding settings in your configuration, but command line Arguments will always win.
+
+Common example:
+
+``` bash
+FREQTRADE__TELEGRAM__CHAT_ID=<telegramchatid>
+FREQTRADE__TELEGRAM__TOKEN=<telegramToken>
+FREQTRADE__EXCHANGE__KEY=<yourExchangeKey>
+FREQTRADE__EXCHANGE__SECRET=<yourExchangeSecret>
+```
+
+Json lists are parsed as json - so you can use the following to set a list of pairs:
+
+``` bash
+export FREQTRADE__EXCHANGE__PAIR_WHITELIST='["BTC/USDT", "ETH/USDT"]'
+```
+
+!!! Note
+    Environment variables detected are logged at startup - so if you can't find why a value is not what you think it should be based on the configuration, make sure it's not loaded from an environment variable.
+
+!!! Tip "Validate combined result"
+    You can use the [show-config subcommand](utils.md#show-config) to see the final, combined configuration.
+
+??? Warning "Loading sequence"
+    Environment variables are loaded after the initial configuration. As such, you cannot provide the path to the configuration through environment variables. Please use `--config path/to/config.json` for that.
+    This also applies to `user_dir` to some degree. while the user directory can be set through environment variables - the configuration will **not** be loaded from that location.
+
+### Multiple configuration files
+
+Multiple configuration files can be specified and used by the bot or the bot can read its configuration parameters from the process standard input stream.
+
+You can specify additional configuration files in `add_config_files`. Files specified in this parameter will be loaded and merged with the initial config file. The files are resolved relative to the initial configuration file.
+This is similar to using multiple `--config` parameters, but simpler in usage as you don't have to specify all files for all commands.
+
+!!! Tip "Validate combined result"
+    You can use the [show-config subcommand](utils.md#show-config) to see the final, combined configuration.
+
+!!! Tip "Use multiple configuration files to keep secrets secret"
+    You can use a 2nd configuration file containing your secrets. That way you can share your "primary" configuration file, while still keeping your API keys for yourself.
+    The 2nd file should only specify what you intend to override.
+    If a key is in more than one of the configurations, then the "last specified configuration" wins (in the above example, `config-private.json`).
+
+    For one-off commands, you can also use the below syntax by specifying multiple "--config" parameters.
+
+    ``` bash
+    freqtrade trade --config user_data/config1.json --config user_data/config-private.json <...>
+    ```
+
+    The below is equivalent to the example above - but having 2 configuration files in the configuration, for easier reuse.
+
+    ``` json title="user_data/config.json"
+    "add_config_files": [
+        "config1.json",
+        "config-private.json"
+    ]
+    ```
+
+    ``` bash
+    freqtrade trade --config user_data/config.json <...>
+    ```
+
+??? Note "config collision handling"
+    If the same configuration setting takes place in both `config.json` and `config-import.json`, then the parent configuration wins.
+    In the below case, `max_open_trades` would be 3 after the merging - as the reusable "import" configuration has this key overwritten.
+
+    ``` json title="user_data/config.json"
+    {
+        "max_open_trades": 3,
+        "stake_currency": "USDT",
+        "add_config_files": [
+            "config-import.json"
+        ]
+    }
+    ```
+
+    ``` json title="user_data/config-import.json"
+    {
+        "max_open_trades": 10,
+        "stake_amount": "unlimited",
+    }
+    ```
+
+    Resulting combined configuration:
+
+    ``` json title="Result"
+    {
+        "max_open_trades": 3,
+        "stake_currency": "USDT",
+        "stake_amount": "unlimited"
+    }
+    ```
+
+    If multiple files are in the `add_config_files` section, then they will be assumed to be at identical levels, having the last occurrence override the earlier config (unless a parent already defined such a key).
+
+## Editor autocomplete and validation
+
+If you are using an editor that supports JSON schema, you can use the schema provided by Freqtrade to get autocompletion and validation of your configuration file by adding the following line to the top of your configuration file:
+
+``` json
+{
+    "$schema": "https://schema.freqtrade.io/schema.json",
+}
+```
+
+??? Note "Develop version"
+    The develop schema is available as `https://schema.freqtrade.io/schema_dev.json` - though we recommend to stick to the stable version for the best experience.
+
+## Configuration parameters
+
+The table below will list all configuration parameters available.
+
+Freqtrade can also load many options via command line (CLI) arguments (check out the commands `--help` output for details).
+
+### Configuration option prevalence
+
+The prevalence for all Options is as follows:
+
+* CLI arguments override any other option
+* [Environment Variables](#environment-variables)
+* Configuration files are used in sequence (the last file wins) and override Strategy configurations.
+* Strategy configurations are only used if they are not set via configuration or command-line arguments. These options are marked with [Strategy Override](#parameters-in-the-strategy) in the below table.
+
+### Parameters table
+
+Mandatory parameters are marked as **Required**, which means that they are required to be set in one of the possible ways.
+
+|  Parameter | Description |
+|------------|-------------|
+| `max_open_trades` | **Required.** Number of open trades your bot is allowed to have. Only one open trade per pair is possible, so the length of your pairlist is another limitation that can apply. If -1 then it is ignored (i.e. potentially unlimited open trades, limited by the pairlist). [More information below](#configuring-amount-per-trade). [Strategy Override](#parameters-in-the-strategy).<br> **Datatype:** Positive integer or -1.
+| `stake_currency` | **Required.** Crypto-currency used for trading. <br> **Datatype:** String
+| `stake_amount` | **Required.** Amount of crypto-currency your bot will use for each trade. Set it to `"unlimited"` to allow the bot to use all available balance. [More information below](#configuring-amount-per-trade). <br> **Datatype:** Positive float or `"unlimited"`.
+| `tradable_balance_ratio` | Ratio of the total account balance the bot is allowed to trade. [More information below](#configuring-amount-per-trade). <br>*Defaults to `0.99` 99%).*<br> **Datatype:** Positive float between `0.1` and `1.0`.
+| `available_capital` | Available starting capital for the bot. Useful when running multiple bots on the same exchange account. [More information below](#configuring-amount-per-trade). <br> **Datatype:** Positive float.
+| `amend_last_stake_amount` | Use reduced last stake amount if necessary. [More information below](#configuring-amount-per-trade). <br>*Defaults to `false`.* <br> **Datatype:** Boolean
+| `last_stake_amount_min_ratio` | Defines minimum stake amount that has to be left and executed. Applies only to the last stake amount when it's amended to a reduced value (i.e. if `amend_last_stake_amount` is set to `true`). [More information below](#configuring-amount-per-trade). <br>*Defaults to `0.5`.* <br> **Datatype:** Float (as ratio)
+| `amount_reserve_percent` | Reserve some amount in min pair stake amount. The bot will reserve `amount_reserve_percent` + stoploss value when calculating min pair stake amount in order to avoid possible trade refusals. <br>*Defaults to `0.05` (5%).* <br> **Datatype:** Positive Float as ratio.
+| `timeframe` | The timeframe to use (e.g `1m`, `5m`, `15m`, `30m`, `1h` ...). Usually missing in configuration, and specified in the strategy. [Strategy Override](#parameters-in-the-strategy). <br> **Datatype:** String
+| `fiat_display_currency` | Fiat currency used to show your profits. [More information below](#what-values-can-be-used-for-fiat_display_currency). <br> **Datatype:** String
+| `dry_run` | **Required.** Define if the bot must be in Dry Run or production mode. <br>*Defaults to `true`.* <br> **Datatype:** Boolean
+| `dry_run_wallet` | Define the starting amount in stake currency for the simulated wallet used by the bot running in Dry Run mode. [More information below](#dry-run-wallet)<br>*Defaults to `1000`.* <br> **Datatype:** Float or Dict
+| `cancel_open_orders_on_exit` | Cancel open orders when the `/stop` RPC command is issued, `Ctrl+C` is pressed or the bot dies unexpectedly. When set to `true`, this allows you to use `/stop` to cancel unfilled and partially filled orders in the event of a market crash. It does not impact open positions. <br>*Defaults to `false`.* <br> **Datatype:** Boolean
+| `process_only_new_candles` | Enable processing of indicators only when new candles arrive. If false each loop populates the indicators, this will mean the same candle is processed many times creating system load but can be useful of your strategy depends on tick data not only candle. [Strategy Override](#parameters-in-the-strategy). <br>*Defaults to `true`.*  <br> **Datatype:** Boolean
+| `minimal_roi` | **Required.** Set the threshold as ratio the bot will use to exit a trade. [More information below](#understand-minimal_roi). [Strategy Override](#parameters-in-the-strategy). <br> **Datatype:** Dict
+| `stoploss` |  **Required.** Value as ratio of the stoploss used by the bot. More details in the [stoploss documentation](stoploss.md). [Strategy Override](#parameters-in-the-strategy).  <br> **Datatype:** Float (as ratio)
+| `trailing_stop` | Enables trailing stoploss (based on `stoploss` in either configuration or strategy file). More details in the [stoploss documentation](stoploss.md#trailing-stop-loss). [Strategy Override](#parameters-in-the-strategy). <br> **Datatype:** Boolean
+| `trailing_stop_positive` | Changes stoploss once profit has been reached. More details in the [stoploss documentation](stoploss.md#trailing-stop-loss-custom-positive-loss). [Strategy Override](#parameters-in-the-strategy). <br> **Datatype:** Float
+| `trailing_stop_positive_offset` | Offset on when to apply `trailing_stop_positive`. Percentage value which should be positive. More details in the [stoploss documentation](stoploss.md#trailing-stop-loss-only-once-the-trade-has-reached-a-certain-offset). [Strategy Override](#parameters-in-the-strategy). <br>*Defaults to `0.0` (no offset).* <br> **Datatype:** Float
+| `trailing_only_offset_is_reached` | Only apply trailing stoploss when the offset is reached. [stoploss documentation](stoploss.md). [Strategy Override](#parameters-in-the-strategy). <br>*Defaults to `false`.*  <br> **Datatype:** Boolean
+| `fee` | Fee used during backtesting / dry-runs. Should normally not be configured, which has freqtrade fall back to the exchange default fee. Set as ratio (e.g. 0.001 = 0.1%). Fee is applied twice for each trade, once when buying, once when selling. <br> **Datatype:** Float (as ratio)
+| `futures_funding_rate` | User-specified funding rate to be used when historical funding rates are not available from the exchange. This does not overwrite real historical rates. It is recommended that this be set to 0 unless you are testing a specific coin and you understand how the funding rate will affect freqtrade's profit calculations. [More information here](leverage.md#unavailable-funding-rates) <br>*Defaults to `None`.*<br> **Datatype:** Float
+| `trading_mode` | Specifies if you want to trade regularly, trade with leverage, or trade contracts whose prices are derived from matching cryptocurrency prices. [leverage documentation](leverage.md). <br>*Defaults to `"spot"`.*  <br> **Datatype:** String
+| `margin_mode` | When trading with leverage, this determines if the collateral owned by the trader will be shared or isolated to each trading pair [leverage documentation](leverage.md). <br> **Datatype:** String
+| `liquidation_buffer` | A ratio specifying how large of a safety net to place between the liquidation price and the stoploss to prevent a position from reaching the liquidation price [leverage documentation](leverage.md). <br>*Defaults to `0.05`.*  <br> **Datatype:** Float
+| | **Unfilled timeout**
+| `unfilledtimeout.entry` | **Required.** How long (in minutes or seconds) the bot will wait for an unfilled entry order to complete, after which the order will be cancelled. [Strategy Override](#parameters-in-the-strategy).<br> **Datatype:** Integer
+| `unfilledtimeout.exit` | **Required.** How long (in minutes or seconds) the bot will wait for an unfilled exit order to complete, after which the order will be cancelled and repeated at current (new) price, as long as there is a signal. [Strategy Override](#parameters-in-the-strategy).<br> **Datatype:** Integer
+| `unfilledtimeout.unit` | Unit to use in unfilledtimeout setting. Note: If you set unfilledtimeout.unit to "seconds", "internals.process_throttle_secs" must be inferior or equal to timeout [Strategy Override](#parameters-in-the-strategy). <br> *Defaults to `"minutes"`.* <br> **Datatype:** String
+| `unfilledtimeout.exit_timeout_count` | How many times can exit orders time out. Once this number of timeouts is reached, an emergency exit is triggered. 0 to disable and allow unlimited order cancels. [Strategy Override](#parameters-in-the-strategy).<br>*Defaults to `0`.* <br> **Datatype:** Integer
+| | **Pricing**
+| `entry_pricing.price_side` | Select the side of the spread the bot should look at to get the entry rate. [More information below](#entry-price).<br> *Defaults to `"same"`.* <br> **Datatype:** String (either `ask`, `bid`, `same` or `other`).
+| `entry_pricing.price_last_balance` | **Required.** Interpolate the bidding price. More information [below](#entry-price-without-orderbook-enabled).
+| `entry_pricing.use_order_book` | Enable entering using the rates in [Order Book Entry](#entry-price-with-orderbook-enabled). <br> *Defaults to `true`.*<br> **Datatype:** Boolean
+| `entry_pricing.order_book_top` | Bot will use the top N rate in Order Book "price_side" to enter a trade. I.e. a value of 2 will allow the bot to pick the 2nd entry in [Order Book Entry](#entry-price-with-orderbook-enabled). <br>*Defaults to `1`.*  <br> **Datatype:** Positive Integer
+| `entry_pricing. check_depth_of_market.enabled` | Do not enter if the difference of buy orders and sell orders is met in Order Book. [Check market depth](#check-depth-of-market). <br>*Defaults to `false`.* <br> **Datatype:** Boolean
+| `entry_pricing. check_depth_of_market.bids_to_ask_delta` | The difference ratio of buy orders and sell orders found in Order Book. A value below 1 means sell order size is greater, while value greater than 1 means buy order size is higher. [Check market depth](#check-depth-of-market) <br> *Defaults to `0`.*  <br> **Datatype:** Float (as ratio)
+| `exit_pricing.price_side` | Select the side of the spread the bot should look at to get the exit rate. [More information below](#exit-price-side).<br> *Defaults to `"same"`.* <br> **Datatype:** String (either `ask`, `bid`, `same` or `other`).
+| `exit_pricing.price_last_balance` | Interpolate the exiting price. More information [below](#exit-price-without-orderbook-enabled).
+| `exit_pricing.use_order_book` | Enable exiting of open trades using [Order Book Exit](#exit-price-with-orderbook-enabled). <br> *Defaults to `true`.*<br> **Datatype:** Boolean
+| `exit_pricing.order_book_top` | Bot will use the top N rate in Order Book "price_side" to exit. I.e. a value of 2 will allow the bot to pick the 2nd ask rate in [Order Book Exit](#exit-price-with-orderbook-enabled)<br>*Defaults to `1`.* <br> **Datatype:** Positive Integer
+| `custom_price_max_distance_ratio` | Configure maximum distance ratio between current and custom entry or exit price. <br>*Defaults to `0.02` 2%).*<br> **Datatype:** Positive float
+| | **TODO**
+| `use_exit_signal` | Use exit signals produced by the strategy in addition to the `minimal_roi`. <br>Setting this to false disables the usage of `"exit_long"` and `"exit_short"` columns. Has no influence on other exit methods (Stoploss, ROI, callbacks). [Strategy Override](#parameters-in-the-strategy). <br>*Defaults to `true`.* <br> **Datatype:** Boolean
+| `exit_profit_only` | Wait until the bot reaches `exit_profit_offset` before taking an exit decision. [Strategy Override](#parameters-in-the-strategy). <br>*Defaults to `false`.* <br> **Datatype:** Boolean
+| `exit_profit_offset` | Exit-signal is only active above this value. Only active in combination with `exit_profit_only=True`. [Strategy Override](#parameters-in-the-strategy). <br>*Defaults to `0.0`.* <br> **Datatype:** Float (as ratio)
+| `ignore_roi_if_entry_signal` | Do not exit if the entry signal is still active. This setting takes preference over `minimal_roi` and `use_exit_signal`. [Strategy Override](#parameters-in-the-strategy). <br>*Defaults to `false`.* <br> **Datatype:** Boolean
+| `ignore_buying_expired_candle_after` | Specifies the number of seconds until a buy signal is no longer used. <br> **Datatype:** Integer
+| `order_types` | Configure order-types depending on the action (`"entry"`, `"exit"`, `"stoploss"`, `"stoploss_on_exchange"`). [More information below](#understand-order_types). [Strategy Override](#parameters-in-the-strategy).<br> **Datatype:** Dict
+| `order_time_in_force` | Configure time in force for entry and exit orders. [More information below](#understand-order_time_in_force). [Strategy Override](#parameters-in-the-strategy). <br> **Datatype:** Dict
+| `position_adjustment_enable` | Enables the strategy to use position adjustments (additional buys or sells). [More information here](strategy-callbacks.md#adjust-trade-position). <br> [Strategy Override](#parameters-in-the-strategy). <br>*Defaults to `false`.*<br> **Datatype:** Boolean
+| `max_entry_position_adjustment` | Maximum additional order(s) for each open trade on top of the first entry Order. Set it to `-1` for unlimited additional orders. [More information here](strategy-callbacks.md#adjust-trade-position). <br> [Strategy Override](#parameters-in-the-strategy). <br>*Defaults to `-1`.*<br> **Datatype:** Positive Integer or -1
+| | **Exchange**
+| `exchange.name` | **Required.** Name of the exchange class to use. <br> **Datatype:** String
+| `exchange.key` | API key to use for the exchange. Only required when you are in production mode.<br>**Keep it in secret, do not disclose publicly.** <br> **Datatype:** String
+| `exchange.secret` | API secret to use for the exchange. Only required when you are in production mode.<br>**Keep it in secret, do not disclose publicly.** <br> **Datatype:** String
+| `exchange.password` | API password to use for the exchange. Only required when you are in production mode and for exchanges that use password for API requests.<br>**Keep it in secret, do not disclose publicly.** <br> **Datatype:** String
+| `exchange.uid` | API uid to use for the exchange. Only required when you are in production mode and for exchanges that use uid for API requests.<br>**Keep it in secret, do not disclose publicly.** <br> **Datatype:** String
+| `exchange.pair_whitelist` | List of pairs to use by the bot for trading and to check for potential trades during backtesting. Supports regex pairs as `.*/BTC`. Not used by VolumePairList. [More information](plugins.md#pairlists-and-pairlist-handlers). <br> **Datatype:** List
+| `exchange.pair_blacklist` | List of pairs the bot must absolutely avoid for trading and backtesting. [More information](plugins.md#pairlists-and-pairlist-handlers). <br> **Datatype:** List
+| `exchange.ccxt_config` | Additional CCXT parameters passed to both ccxt instances (sync and async). This is usually the correct place for additional ccxt configurations. Parameters may differ from exchange to exchange and are documented in the [ccxt documentation](https://docs.ccxt.com/#/README?id=overriding-exchange-properties-upon-instantiation). Please avoid adding exchange secrets here (use the dedicated fields instead), as they may be contained in logs. <br> **Datatype:** Dict
+| `exchange.ccxt_sync_config` | Additional CCXT parameters passed to the regular (sync) ccxt instance. Parameters may differ from exchange to exchange and are documented in the [ccxt documentation](https://docs.ccxt.com/#/README?id=overriding-exchange-properties-upon-instantiation) <br> **Datatype:** Dict
+| `exchange.ccxt_async_config` | Additional CCXT parameters passed to the async ccxt instance. Parameters may differ from exchange to exchange  and are documented in the [ccxt documentation](https://docs.ccxt.com/#/README?id=overriding-exchange-properties-upon-instantiation) <br> **Datatype:** Dict
+| `exchange.enable_ws` | Enable the usage of Websockets for the exchange. <br>[More information](#consuming-exchange-websockets).<br>*Defaults to `true`.* <br> **Datatype:** Boolean
+| `exchange.markets_refresh_interval` | The interval in minutes in which markets are reloaded. <br>*Defaults to `60` minutes.* <br> **Datatype:** Positive Integer
+| `exchange.skip_open_order_update` | Skips open order updates on startup should the exchange cause problems. Only relevant in live conditions.<br>*Defaults to `false`*<br> **Datatype:** Boolean
+| `exchange.unknown_fee_rate` | Fallback value to use when calculating trading fees. This can be useful for exchanges which have fees in non-tradable currencies. The value provided here will be multiplied with the "fee cost".<br>*Defaults to `None`<br> **Datatype:** float
+| `exchange.log_responses` | Log relevant exchange responses. For debug mode only - use with care.<br>*Defaults to `false`*<br> **Datatype:** Boolean
+| `exchange.only_from_ccxt` | Prevent data-download from data.binance.vision. Leaving this as false can greatly speed up downloads, but may be problematic if the site is not available.<br>*Defaults to `false`*<br> **Datatype:** Boolean
+| `experimental.block_bad_exchanges` | Block exchanges known to not work with freqtrade. Leave on default unless you want to test if that exchange works now. <br>*Defaults to `true`.* <br> **Datatype:** Boolean
+| | **Plugins**
+| `edge.*` | Please refer to [edge configuration document](edge.md) for detailed explanation of all possible configuration options.
+| `pairlists` | Define one or more pairlists to be used. [More information](plugins.md#pairlists-and-pairlist-handlers). <br>*Defaults to `StaticPairList`.*  <br> **Datatype:** List of Dicts
+| | **Telegram**
+| `telegram.enabled` | Enable the usage of Telegram. <br> **Datatype:** Boolean
+| `telegram.token` | Your Telegram bot token. Only required if `telegram.enabled` is `true`. <br>**Keep it in secret, do not disclose publicly.** <br> **Datatype:** String
+| `telegram.chat_id` | Your personal Telegram account id. Only required if `telegram.enabled` is `true`. <br>**Keep it in secret, do not disclose publicly.** <br> **Datatype:** String
+| `telegram.balance_dust_level` | Dust-level (in stake currency) - currencies with a balance below this will not be shown by `/balance`. <br> **Datatype:** float
+| `telegram.reload` | Allow "reload" buttons on telegram messages. <br>*Defaults to `true`.<br> **Datatype:** boolean
+| `telegram.notification_settings.*` | Detailed notification settings. Refer to the [telegram documentation](telegram-usage.md) for details.<br> **Datatype:** dictionary
+| `telegram.allow_custom_messages` | Enable the sending of Telegram messages from strategies via the dataprovider.send_msg() function. <br> **Datatype:** Boolean
+| | **Webhook**
+| `webhook.enabled` | Enable usage of Webhook notifications <br> **Datatype:** Boolean
+| `webhook.url` | URL for the webhook. Only required if `webhook.enabled` is `true`. See the [webhook documentation](webhook-config.md) for more details. <br> **Datatype:** String
+| `webhook.entry` | Payload to send on entry. Only required if `webhook.enabled` is `true`. See the [webhook documentation](webhook-config.md) for more details. <br> **Datatype:** String
+| `webhook.entry_cancel` | Payload to send on entry order cancel. Only required if `webhook.enabled` is `true`. See the [webhook documentation](webhook-config.md) for more details. <br> **Datatype:** String
+| `webhook.entry_fill` | Payload to send on entry order filled. Only required if `webhook.enabled` is `true`. See the [webhook documentation](webhook-config.md) for more details. <br> **Datatype:** String
+| `webhook.exit` | Payload to send on exit. Only required if `webhook.enabled` is `true`. See the [webhook documentation](webhook-config.md) for more details. <br> **Datatype:** String
+| `webhook.exit_cancel` | Payload to send on exit order cancel. Only required if `webhook.enabled` is `true`. See the [webhook documentation](webhook-config.md) for more details. <br> **Datatype:** String
+| `webhook.exit_fill` | Payload to send on exit order filled. Only required if `webhook.enabled` is `true`. See the [webhook documentation](webhook-config.md) for more details. <br> **Datatype:** String
+| `webhook.status` | Payload to send on status calls. Only required if `webhook.enabled` is `true`. See the [webhook documentation](webhook-config.md) for more details. <br> **Datatype:** String
+| `webhook.allow_custom_messages` | Enable the sending of Webhook messages from strategies via the dataprovider.send_msg() function. <br> **Datatype:** Boolean
+| | **Rest API / FreqUI / Producer-Consumer**
+| `api_server.enabled` | Enable usage of API Server. See the [API Server documentation](rest-api.md) for more details. <br> **Datatype:** Boolean
+| `api_server.listen_ip_address` | Bind IP address. See the [API Server documentation](rest-api.md) for more details. <br> **Datatype:** IPv4
+| `api_server.listen_port` | Bind Port. See the [API Server documentation](rest-api.md) for more details. <br>**Datatype:** Integer between 1024 and 65535
+| `api_server.verbosity` | Logging verbosity. `info` will print all RPC Calls, while "error" will only display errors. <br>**Datatype:** Enum, either `info` or `error`. Defaults to `info`.
+| `api_server.username` | Username for API server. See the [API Server documentation](rest-api.md) for more details. <br>**Keep it in secret, do not disclose publicly.**<br> **Datatype:** String
+| `api_server.password` | Password for API server. See the [API Server documentation](rest-api.md) for more details. <br>**Keep it in secret, do not disclose publicly.**<br> **Datatype:** String
+| `api_server.ws_token` | API token for the Message WebSocket. See the [API Server documentation](rest-api.md) for more details.  <br>**Keep it in secret, do not disclose publicly.** <br> **Datatype:** String
+| `bot_name` | Name of the bot. Passed via API to a client - can be shown to distinguish / name bots.<br> *Defaults to `freqtrade`*<br> **Datatype:** String
+| `external_message_consumer` | Enable [Producer/Consumer mode](producer-consumer.md) for more details. <br> **Datatype:** Dict
+| | **Other**
+| `initial_state` | Defines the initial application state. If set to stopped, then the bot has to be explicitly started via `/start` RPC command. <br>*Defaults to `stopped`.* <br> **Datatype:** Enum, either `running`, `paused` or `stopped`
+| `force_entry_enable` | Enables the RPC Commands to force a Trade entry. More information below. <br> **Datatype:** Boolean
+| `disable_dataframe_checks` | Disable checking the OHLCV dataframe returned from the strategy methods for correctness. Only use when intentionally changing the dataframe and understand what you are doing. [Strategy Override](#parameters-in-the-strategy).<br> *Defaults to `False`*. <br> **Datatype:** Boolean
+| `internals.process_throttle_secs` | Set the process throttle, or minimum loop duration for one bot iteration loop. Value in second. <br>*Defaults to `5` seconds.* <br> **Datatype:** Positive Integer
+| `internals.heartbeat_interval` | Print heartbeat message every N seconds. Set to 0 to disable heartbeat messages. <br>*Defaults to `60` seconds.* <br> **Datatype:** Positive Integer or 0
+| `internals.sd_notify` | Enables use of the sd_notify protocol to tell systemd service manager about changes in the bot state and issue keep-alive pings. See [here](advanced-setup.md#configure-the-bot-running-as-a-systemd-service) for more details. <br> **Datatype:** Boolean
+| `strategy` | **Required** Defines Strategy class to use. Recommended to be set via `--strategy NAME`. <br> **Datatype:** ClassName
+| `strategy_path` | Adds an additional strategy lookup path (must be a directory). <br> **Datatype:** String
+| `recursive_strategy_search` | Set to `true` to recursively search sub-directories inside `user_data/strategies` for a strategy. <br> **Datatype:** Boolean
+| `user_data_dir` | Directory containing user data. <br> *Defaults to `./user_data/`*. <br> **Datatype:** String
+| `db_url` | Declares database URL to use. NOTE: This defaults to `sqlite:///tradesv3.dryrun.sqlite` if `dry_run` is `true`, and to `sqlite:///tradesv3.sqlite` for production instances. <br> **Datatype:** String, SQLAlchemy connect string
+| `logfile` | Specifies logfile name. Uses a rolling strategy for log file rotation for 10 files with the 1MB limit per file. <br> **Datatype:** String
+| `add_config_files` | Additional config files. These files will be loaded and merged with the current config file. The files are resolved relative to the initial file.<br> *Defaults to `[]`*. <br> **Datatype:** List of strings
+| `dataformat_ohlcv` | Data format to use to store historical candle (OHLCV) data. <br> *Defaults to `feather`*. <br> **Datatype:** String
+| `dataformat_trades` | Data format to use to store historical trades data. <br> *Defaults to `feather`*. <br> **Datatype:** String
+| `reduce_df_footprint` | Recast all numeric columns to float32/int32, with the objective of reducing ram/disk usage (and decreasing train/inference timing in FreqAI). (Currently only affects FreqAI use-cases) <br> **Datatype:** Boolean. <br> Default: `False`.
+| `log_config` | Dictionary containing the log config for python logging. [more info](advanced-setup.md#advanced-logging) <br> **Datatype:** dict. <br> Default: `FtRichHandler`
+
+### Parameters in the strategy
+
+The following parameters can be set in the configuration file or strategy.
+Values set in the configuration file always overwrite values set in the strategy.
+
+* `minimal_roi`
+* `timeframe`
+* `stoploss`
+* `max_open_trades`
+* `trailing_stop`
+* `trailing_stop_positive`
+* `trailing_stop_positive_offset`
+* `trailing_only_offset_is_reached`
+* `use_custom_stoploss`
+* `process_only_new_candles`
+* `order_types`
+* `order_time_in_force`
+* `unfilledtimeout`
+* `disable_dataframe_checks`
+* `use_exit_signal`
+* `exit_profit_only`
+* `exit_profit_offset`
+* `ignore_roi_if_entry_signal`
+* `ignore_buying_expired_candle_after`
+* `position_adjustment_enable`
+* `max_entry_position_adjustment`
+
+### Configuring amount per trade
+
+There are several methods to configure how much of the stake currency the bot will use to enter a trade. All methods respect the [available balance configuration](#tradable-balance) as explained below.
+
+#### Minimum trade stake
+
+The minimum stake amount will depend on exchange and pair and is usually listed in the exchange support pages.
+
+Assuming the minimum tradable amount for XRP/USD is 20 XRP (given by the exchange), and the price is 0.6\$, the minimum stake amount to buy this pair is `20 * 0.6 ~= 12`.
+This exchange has also a limit on USD - where all orders must be > 10\$ - which however does not apply in this case.
+
+To guarantee safe execution, freqtrade will not allow buying with a stake-amount of 10.1\$, instead, it'll make sure that there's enough space to place a stoploss below the pair (+ an offset, defined by `amount_reserve_percent`, which defaults to 5%).
+
+With a reserve of 5%, the minimum stake amount would be ~12.6\$ (`12 * (1 + 0.05)`). If we take into account a stoploss of 10% on top of that - we'd end up with a value of ~14\$ (`12.6 / (1 - 0.1)`).
+
+To limit this calculation in case of large stoploss values, the calculated minimum stake-limit will never be more than 50% above the real limit.
+
+!!! Warning
+    Since the limits on exchanges are usually stable and are not updated often, some pairs can show pretty high minimum limits, simply because the price increased a lot since the last limit adjustment by the exchange. Freqtrade adjusts the stake-amount to this value, unless it's > 30% more than the calculated/desired stake-amount - in which case the trade is rejected.
+
+#### Dry-run wallet
+
+When running in dry-run mode, the bot will use a simulated wallet to execute trades. The starting balance of this wallet is defined by `dry_run_wallet` (defaults to 1000).
+For more complex scenarios, you can also assign a dictionary to `dry_run_wallet` to define the starting balance for each currency.
+
+```json
+"dry_run_wallet": {
+    "BTC": 0.01,
+    "ETH": 2,
+    "USDT": 1000
+}
+```
+
+Command line options (`--dry-run-wallet`) can be used to override the configuration value, but only for the float value, not for the dictionary. If you'd like to use the dictionary, please adjust the configuration file.
+
+!!! Note
+    Balances not in stake-currency will not be used for trading, but are shown as part of the wallet balance.
+    On Cross-margin exchanges, the wallet balance may be used to calculate the available collateral for trading.
+
+#### Tradable balance
+
+By default, the bot assumes that the `complete amount - 1%` is at it's disposal, and when using [dynamic stake amount](#dynamic-stake-amount), it will split the complete balance into `max_open_trades` buckets per trade.
+Freqtrade will reserve 1% for eventual fees when entering a trade and will therefore not touch that by default.
+
+You can configure the "untouched" amount by using the `tradable_balance_ratio` setting.
+
+For example, if you have 10 ETH available in your wallet on the exchange and `tradable_balance_ratio=0.5` (which is 50%), then the bot will use a maximum amount of 5 ETH for trading and considers this as an available balance. The rest of the wallet is untouched by the trades.
+
+!!! Danger
+    This setting should **not** be used when running multiple bots on the same account. Please look at [Available Capital to the bot](#assign-available-capital) instead.
+
+!!! Warning
+    The `tradable_balance_ratio` setting applies to the current balance (free balance + tied up in trades). Therefore, assuming the starting balance of 1000, a configuration with `tradable_balance_ratio=0.99` will not guarantee that 10 currency units will always remain available on the exchange. For example, the free amount may reduce to 5 units if the total balance is reduced to 500 (either by a losing streak or by withdrawing balance).
+
+#### Assign available Capital
+
+To fully utilize compounding profits when using multiple bots on the same exchange account, you'll want to limit each bot to a certain starting balance.
+This can be accomplished by setting `available_capital` to the desired starting balance.
+
+Assuming your account has 10000 USDT and you want to run 2 different strategies on this exchange.
+You'd set `available_capital=5000` - granting each bot an initial capital of 5000 USDT.
+The bot will then split this starting balance equally into `max_open_trades` buckets.
+Profitable trades will result in increased stake-sizes for this bot - without affecting the stake-sizes of the other bot.
+
+Adjusting `available_capital` requires reloading the configuration to take effect. Adjusting the `available_capital` adds the difference between the previous `available_capital` and the new `available_capital`. Decreasing the available capital when trades are open doesn't exit the trades. The difference is returned to the wallet when the trades conclude. The outcome of this differs depending on the price movement between the adjustment and exiting the trades.
+
+!!! Warning "Incompatible with `tradable_balance_ratio`"
+    Setting this option will replace any configuration of `tradable_balance_ratio`.
+
+#### Amend last stake amount
+
+Assuming we have the tradable balance of 1000 USDT, `stake_amount=400`, and `max_open_trades=3`.
+The bot would open 2 trades and will be unable to fill the last trading slot, since the requested 400 USDT are no longer available since 800 USDT are already tied in other trades.
+
+To overcome this, the option `amend_last_stake_amount` can be set to `True`, which will enable the bot to reduce stake_amount to the available balance to fill the last trade slot.
+
+In the example above this would mean:
+
+* Trade1: 400 USDT
+* Trade2: 400 USDT
+* Trade3: 200 USDT
+
+!!! Note
+    This option only applies with [Static stake amount](#static-stake-amount) - since [Dynamic stake amount](#dynamic-stake-amount) divides the balances evenly.
+
+!!! Note
+    The minimum last stake amount can be configured using `last_stake_amount_min_ratio` - which defaults to 0.5 (50%). This means that the minimum stake amount that's ever used is `stake_amount * 0.5`. This avoids very low stake amounts, that are close to the minimum tradable amount for the pair and can be refused by the exchange.
+
+#### Static stake amount
+
+The `stake_amount` configuration statically configures the amount of stake-currency your bot will use for each trade.
+
+The minimal configuration value is 0.0001, however, please check your exchange's trading minimums for the stake currency you're using to avoid problems.
+
+This setting works in combination with `max_open_trades`. The maximum capital engaged in trades is `stake_amount * max_open_trades`.
+For example, the bot will at most use (0.05 BTC x 3) = 0.15 BTC, assuming a configuration of `max_open_trades=3` and `stake_amount=0.05`.
+
+!!! Note
+    This setting respects the [available balance configuration](#tradable-balance).
+
+#### Dynamic stake amount
+
+Alternatively, you can use a dynamic stake amount, which will use the available balance on the exchange, and divide that equally by the number of allowed trades (`max_open_trades`).
+
+To configure this, set `stake_amount="unlimited"`. We also recommend to set `tradable_balance_ratio=0.99` (99%) - to keep a minimum balance for eventual fees.
+
+In this case a trade amount is calculated as:
+
+```python
+currency_balance / (max_open_trades - current_open_trades)
+```
+
+To allow the bot to trade all the available `stake_currency` in your account (minus `tradable_balance_ratio`) set
+
+```json
+"stake_amount" : "unlimited",
+"tradable_balance_ratio": 0.99,
+```
+
+!!! Tip "Compounding profits"
+    This configuration will allow increasing/decreasing stakes depending on the performance of the bot (lower stake if the bot is losing, higher stakes if the bot has a winning record since higher balances are available), and will result in profit compounding.
+
+!!! Note "When using Dry-Run Mode"
+    When using `"stake_amount" : "unlimited",` in combination with Dry-Run, Backtesting or Hyperopt, the balance will be simulated starting with a stake of `dry_run_wallet` which will evolve.
+    It is therefore important to set `dry_run_wallet` to a sensible value (like 0.05 or 0.01 for BTC and 1000 or 100 for USDT, for example), otherwise, it may simulate trades with 100 BTC (or more) or 0.05 USDT (or less) at once - which may not correspond to your real available balance or is less than the exchange minimal limit for the order amount for the stake currency.
+
+#### Dynamic stake amount with position adjustment
+
+When you want to use position adjustment with unlimited stakes, you must also implement `custom_stake_amount` to a return a value depending on your strategy.
+Typical value would be in the range of 25% - 50% of the proposed stakes, but depends highly on your strategy and how much you wish to leave into the wallet as position adjustment buffer.
+
+For example if your position adjustment assumes it can do 2 additional buys with the same stake amounts then your buffer should be 66.6667% of the initially proposed unlimited stake amount.
+
+Or another example if your position adjustment assumes it can do 1 additional buy with 3x the original stake amount then `custom_stake_amount` should return 25% of proposed stake amount and leave 75% for possible later position adjustments.
+
+--8<-- "includes/pricing.md"
+
+## Further Configuration details
+
+### Understand minimal_roi
+
+The `minimal_roi` configuration parameter is a JSON object where the key is a duration
+in minutes and the value is the minimum ROI as a ratio.
+See the example below:
+
+```json
+"minimal_roi": {
+    "40": 0.0,    # Exit after 40 minutes if the profit is not negative
+    "30": 0.01,   # Exit after 30 minutes if there is at least 1% profit
+    "20": 0.02,   # Exit after 20 minutes if there is at least 2% profit
+    "0":  0.04    # Exit immediately if there is at least 4% profit
+},
+```
+
+Most of the strategy files already include the optimal `minimal_roi` value.
+This parameter can be set in either Strategy or Configuration file. If you use it in the configuration file, it will override the
+`minimal_roi` value from the strategy file.
+If it is not set in either Strategy or Configuration, a default of 1000% `{"0": 10}` is used, and minimal ROI is disabled unless your trade generates 1000% profit.
+
+!!! Note "Special case to forceexit after a specific time"
+    A special case presents using `"<N>": -1` as ROI. This forces the bot to exit a trade after N Minutes, no matter if it's positive or negative, so represents a time-limited force-exit.
+
+### Understand force_entry_enable
+
+The `force_entry_enable` configuration parameter enables the usage of force-enter (`/forcelong`, `/forceshort`) commands via Telegram and REST API.
+For security reasons, it's disabled by default, and freqtrade will show a warning message on startup if enabled.
+For example, you can send `/forceenter ETH/BTC` to the bot, which will result in freqtrade buying the pair and holds it until a regular exit-signal (ROI, stoploss, /forceexit) appears.
+
+This can be dangerous with some strategies, so use with care.
+
+See [the telegram documentation](telegram-usage.md) for details on usage.
+
+### Ignoring expired candles
+
+When working with larger timeframes (for example 1h or more) and using a low `max_open_trades` value, the last candle can be processed as soon as a trade slot becomes available. When processing the last candle, this can lead to a situation where it may not be desirable to use the buy signal on that candle. For example, when using a condition in your strategy where you use a cross-over, that point may have passed too long ago for you to start a trade on it.
+
+In these situations, you can enable the functionality to ignore candles that are beyond a specified period by setting `ignore_buying_expired_candle_after` to a positive number, indicating the number of seconds after which the buy signal becomes expired.
+
+For example, if your strategy is using a 1h timeframe, and you only want to buy within the first 5 minutes when a new candle comes in, you can add the following configuration to your strategy:
+
+``` json
+  {
+    //...
+    "ignore_buying_expired_candle_after": 300,
+    // ...
+  }
+```
+
+!!! Note
+    This setting resets with each new candle, so it will not prevent sticking-signals from executing on the 2nd or 3rd candle they're active. Best use a "trigger" selector for buy signals, which are only active for one candle.
+
+### Understand order_types
+
+The `order_types` configuration parameter maps actions (`entry`, `exit`, `stoploss`, `emergency_exit`, `force_exit`, `force_entry`) to order-types (`market`, `limit`, ...) as well as configures stoploss to be on the exchange and defines stoploss on exchange update interval in seconds.
+
+This allows to enter using limit orders, exit using limit-orders, and create stoplosses using market orders.
+It also allows to set the
+stoploss "on exchange" which means stoploss order would be placed immediately once the buy order is fulfilled.
+
+`order_types` set in the configuration file overwrites values set in the strategy as a whole, so you need to configure the whole `order_types` dictionary in one place.
+
+If this is configured, the following 4 values (`entry`, `exit`, `stoploss` and `stoploss_on_exchange`) need to be present, otherwise, the bot will fail to start.
+
+For information on (`emergency_exit`,`force_exit`, `force_entry`, `stoploss_on_exchange`,`stoploss_on_exchange_interval`,`stoploss_on_exchange_limit_ratio`) please see stop loss documentation [stop loss on exchange](stoploss.md)
+
+Syntax for Strategy:
+
+```python
+order_types = {
+    "entry": "limit",
+    "exit": "limit",
+    "emergency_exit": "market",
+    "force_entry": "market",
+    "force_exit": "market",
+    "stoploss": "market",
+    "stoploss_on_exchange": False,
+    "stoploss_on_exchange_interval": 60,
+    "stoploss_on_exchange_limit_ratio": 0.99,
+}
+```
+
+Configuration:
+
+```json
+"order_types": {
+    "entry": "limit",
+    "exit": "limit",
+    "emergency_exit": "market",
+    "force_entry": "market",
+    "force_exit": "market",
+    "stoploss": "market",
+    "stoploss_on_exchange": false,
+    "stoploss_on_exchange_interval": 60
+}
+```
+
+!!! Note "Market order support"
+    Not all exchanges support "market" orders.
+    The following message will be shown if your exchange does not support market orders:
+    `"Exchange <yourexchange> does not support market orders."` and the bot will refuse to start.
+
+!!! Warning "Using market orders"
+    Please carefully read the section [Market order pricing](#market-order-pricing) section when using market orders.
+
+!!! Note "Stoploss on exchange"
+    `order_types.stoploss_on_exchange_interval` is not mandatory. Do not change its value if you are
+    unsure of what you are doing. For more information about how stoploss works please
+    refer to [the stoploss documentation](stoploss.md).
+
+    If `order_types.stoploss_on_exchange` is enabled and the stoploss is cancelled manually on the exchange, then the bot will create a new stoploss order.
+
+!!! Warning "Warning: order_types.stoploss_on_exchange failures"
+    If stoploss on exchange creation fails for some reason, then an "emergency exit" is initiated. By default, this will exit the trade using a market order. The order-type for the emergency-exit can be changed by setting the `emergency_exit` value in the `order_types` dictionary - however, this is not advised.
+
+### Understand order_time_in_force
+
+The `order_time_in_force` configuration parameter defines the policy by which the order
+is executed on the exchange. Three commonly used time in force are:
+
+**GTC (Good Till Canceled):**
+
+This is most of the time the default time in force. It means the order will remain
+on exchange till it is cancelled by the user. It can be fully or partially fulfilled.
+If partially fulfilled, the remaining will stay on the exchange till cancelled.
+
+**FOK (Fill Or Kill):**
+
+It means if the order is not executed immediately AND fully then it is cancelled by the exchange.
+
+**IOC (Immediate Or Canceled):**
+
+It is the same as FOK (above) except it can be partially fulfilled. The remaining part
+is automatically cancelled by the exchange.
+
+**PO (Post only):**
+
+Post only order. The order is either placed as a maker order, or it is canceled.
+This means the order must be placed on orderbook for at least time in an unfilled state.
+
+#### time_in_force config
+
+The `order_time_in_force` parameter contains a dict with entry and exit time in force policy values.
+This can be set in the configuration file or in the strategy.
+Values set in the configuration file overwrites values set in the strategy.
+
+The possible values are: `GTC` (default), `FOK` or `IOC`.
+
+``` python
+"order_time_in_force": {
+    "entry": "GTC",
+    "exit": "GTC"
+},
+```
+
+!!! Warning
+    This is ongoing work. For now, it is supported only for binance, gate and kucoin.
+    Please don't change the default value unless you know what you are doing and have researched the impact of using different values for your particular exchange.
+
+### Fiat conversion
+
+Freqtrade uses the Coingecko API to convert the coin value to it's corresponding fiat value for the Telegram reports.
+The FIAT currency can be set in the configuration file as `fiat_display_currency`.
+
+Removing `fiat_display_currency` completely from the configuration will skip initializing coingecko, and will not show any FIAT currency conversion. This has no importance for the correct functioning of the bot.
+
+#### What values can be used for fiat_display_currency?
+
+The `fiat_display_currency` configuration parameter sets the base currency to use for the
+conversion from coin to fiat in the bot Telegram reports.
+
+The valid values are:
+
+```json
+"AUD", "BRL", "CAD", "CHF", "CLP", "CNY", "CZK", "DKK", "EUR", "GBP", "HKD", "HUF", "IDR", "ILS", "INR", "JPY", "KRW", "MXN", "MYR", "NOK", "NZD", "PHP", "PKR", "PLN", "RUB", "SEK", "SGD", "THB", "TRY", "TWD", "ZAR", "USD"
+```
+
+In addition to fiat currencies, a range of crypto currencies is supported.
+
+The valid values are:
+
+```json
+"BTC", "ETH", "XRP", "LTC", "BCH", "BNB"
+```
+
+#### Coingecko Rate limit problems
+
+On some IP ranges, coingecko is heavily rate-limiting.
+In such cases, you may want to add your coingecko API key to the configuration.
+
+``` json
+{
+    "fiat_display_currency": "USD",
+    "coingecko": {
+        "api_key": "your-api",
+        "is_demo": true
+    }
+}
+```
+
+Freqtrade supports both Demo and Pro coingecko API keys.
+
+The Coingecko API key is NOT required for the bot to function correctly.
+It is only used for the conversion of coin to fiat in the Telegram reports, which usually also work without API key.
+
+## Consuming exchange Websockets
+
+Freqtrade can consume websockets through ccxt.pro.
+
+Freqtrade aims ensure data is available at all times.
+Should the websocket connection fail (or be disabled), the bot will fall back to REST API calls.
+
+Should you experience problems you suspect are caused by websockets, you can disable these via the setting `exchange.enable_ws`, which defaults to true.
+
+```jsonc
+"exchange": {
+    // ...
+    "enable_ws": false,
+    // ...
+}
+```
+
+Should you be required to use a proxy, please refer to the [proxy section](#using-proxy-with-freqtrade) for more information.
+
+!!! Info "Rollout"
+    We're implementing this out slowly, ensuring stability of your bots.
+    Currently, usage is limited to ohlcv data streams.
+    It's also limited to a few exchanges, with new exchanges being added on an ongoing basis.
+
+## Using Dry-run mode
+
+We recommend starting the bot in the Dry-run mode to see how your bot will
+behave and what is the performance of your strategy. In the Dry-run mode, the
+bot does not engage your money. It only runs a live simulation without
+creating trades on the exchange.
+
+1. Edit your `config.json` configuration file.
+2. Switch `dry-run` to `true` and specify `db_url` for a persistence database.
+
+```json
+"dry_run": true,
+"db_url": "sqlite:///tradesv3.dryrun.sqlite",
+```
+
+3. Remove your Exchange API key and secret (change them by empty values or fake credentials):
+
+```json
+"exchange": {
+    "name": "binance",
+    "key": "key",
+    "secret": "secret",
+    ...
+}
+```
+
+Once you will be happy with your bot performance running in the Dry-run mode, you can switch it to production mode.
+
+!!! Note
+    A simulated wallet is available during dry-run mode and will assume a starting capital of `dry_run_wallet` (defaults to 1000).
+
+### Considerations for dry-run
+
+* API-keys may or may not be provided. Only Read-Only operations (i.e. operations that do not alter account state) on the exchange are performed in dry-run mode.
+* Wallets (`/balance`) are simulated based on `dry_run_wallet`.
+* Orders are simulated, and will not be posted to the exchange.
+* Market orders fill based on orderbook volume the moment the order is placed, with a maximum slippage of 5%.
+* Limit orders fill once the price reaches the defined level - or time out based on `unfilledtimeout` settings.
+* Limit orders will be converted to market orders if they cross the price by more than 1%, and will be filled immediately based regular market order rules (see point about Market orders above).
+* In combination with `stoploss_on_exchange`, the stop_loss price is assumed to be filled.
+* Open orders (not trades, which are stored in the database) are kept open after bot restarts, with the assumption that they were not filled while being offline.
+
+## Switch to production mode
+
+In production mode, the bot will engage your money. Be careful, since a wrong strategy can lose all your money.
+Be aware of what you are doing when you run it in production mode.
+
+When switching to Production mode, please make sure to use a different / fresh database to avoid dry-run trades messing with your exchange money and eventually tainting your statistics.
+
+### Setup your exchange account
+
+You will need to create API Keys (usually you get `key` and `secret`, some exchanges require an additional `password`) from the Exchange website and you'll need to insert this into the appropriate fields in the configuration or when asked by the `freqtrade new-config` command.
+API Keys are usually only required for live trading (trading for real money, bot running in "production mode", executing real orders on the exchange) and are not required for the bot running in dry-run (trade simulation) mode. When you set up the bot in dry-run mode, you may fill these fields with empty values.
+
+### To switch your bot in production mode
+
+**Edit your `config.json`  file.**
+
+**Switch dry-run to false and don't forget to adapt your database URL if set:**
+
+```json
+"dry_run": false,
+```
+
+**Insert your Exchange API key (change them by fake API keys):**
+
+```json
+{
+    "exchange": {
+        "name": "binance",
+        "key": "af8ddd35195e9dc500b9a6f799f6f5c93d89193b",
+        "secret": "08a9dc6db3d7b53e1acebd9275677f4b0a04f1a5",
+        //"password": "", // Optional, not needed by all exchanges)
+        // ...
+    }
+    //...
+}
+```
+
+You should also make sure to read the [Exchanges](exchanges.md) section of the documentation to be aware of potential configuration details specific to your exchange.
+
+!!! Hint "Keep your secrets secret"
+    To keep your secrets secret, we recommend using a 2nd configuration for your API keys.
+    Simply use the above snippet in a new configuration file (e.g. `config-private.json`) and keep your settings in this file.
+    You can then start the bot with `freqtrade trade --config user_data/config.json --config user_data/config-private.json <...>` to have your keys loaded.
+
+    **NEVER** share your private configuration file or your exchange keys with anyone!
+
+## Using a proxy with Freqtrade
+
+To use a proxy with freqtrade, export your proxy settings using the variables `"HTTP_PROXY"` and `"HTTPS_PROXY"` set to the appropriate values.
+This will have the proxy settings applied to everything (telegram, coingecko, ...) **except** for exchange requests.
+
+``` bash
+export HTTP_PROXY="http://addr:port"
+export HTTPS_PROXY="http://addr:port"
+freqtrade
+```
+
+### Proxy exchange requests
+
+To use a proxy for exchange connections - you will have to define the proxies as part of the ccxt configuration.
+
+``` json
+{ 
+  "exchange": {
+    "ccxt_config": {
+      "httpsProxy": "http://addr:port",
+      "wsProxy": "http://addr:port",
+    }
+  }
+}
+```
+
+For more information on available proxy types, please consult the [ccxt proxy documentation](https://docs.ccxt.com/#/README?id=proxy).
+
+## Next step
+
+Now you have configured your config.json, the next step is to [start your bot](bot-usage.md).

md/freqai-configuration.md

@@ -0,0 +1,423 @@
+# Configuration
+
+FreqAI is configured through the typical [Freqtrade config file](configuration.md) and the standard [Freqtrade strategy](strategy-customization.md). Examples of FreqAI config and strategy files can be found in `config_examples/config_freqai.example.json` and `freqtrade/templates/FreqaiExampleStrategy.py`, respectively.
+
+## Setting up the configuration file
+
+ Although there are plenty of additional parameters to choose from, as highlighted in the [parameter table](freqai-parameter-table.md#parameter-table), a FreqAI config must at minimum include the following parameters (the parameter values are only examples):
+
+```json
+    "freqai": {
+        "enabled": true,
+        "purge_old_models": 2,
+        "train_period_days": 30,
+        "backtest_period_days": 7,
+        "identifier" : "unique-id",
+        "feature_parameters" : {
+            "include_timeframes": ["5m","15m","4h"],
+            "include_corr_pairlist": [
+                "ETH/USD",
+                "LINK/USD",
+                "BNB/USD"
+            ],
+            "label_period_candles": 24,
+            "include_shifted_candles": 2,
+            "indicator_periods_candles": [10, 20]
+        },
+        "data_split_parameters" : {
+            "test_size": 0.25
+        }
+    }
+```
+
+A full example config is available in `config_examples/config_freqai.example.json`.
+
+!!! Note
+    The `identifier` is commonly overlooked by newcomers, however, this value plays an important role in your configuration. This value is a unique ID that you choose to describe one of your runs. Keeping it the same allows you to maintain crash resilience as well as faster backtesting. As soon as you want to try a new run (new features, new model, etc.), you should change this value (or delete the `user_data/models/unique-id` folder. More details available in the [parameter table](freqai-parameter-table.md#feature-parameters).
+
+## Building a FreqAI strategy
+
+The FreqAI strategy requires including the following lines of code in the standard [Freqtrade strategy](strategy-customization.md):
+
+```python
+    # user should define the maximum startup candle count (the largest number of candles
+    # passed to any single indicator)
+    startup_candle_count: int = 20
+
+    def populate_indicators(self, dataframe: DataFrame, metadata: dict) -> DataFrame:
+
+        # the model will return all labels created by user in `set_freqai_targets()`
+        # (& appended targets), an indication of whether or not the prediction should be accepted,
+        # the target mean/std values for each of the labels created by user in
+        # `set_freqai_targets()` for each training period.
+
+        dataframe = self.freqai.start(dataframe, metadata, self)
+
+        return dataframe
+
+    def feature_engineering_expand_all(self, dataframe: DataFrame, period, **kwargs) -> DataFrame:
+        """
+        *Only functional with FreqAI enabled strategies*
+        This function will automatically expand the defined features on the config defined
+        `indicator_periods_candles`, `include_timeframes`, `include_shifted_candles`, and
+        `include_corr_pairs`. In other words, a single feature defined in this function
+        will automatically expand to a total of
+        `indicator_periods_candles` * `include_timeframes` * `include_shifted_candles` *
+        `include_corr_pairs` numbers of features added to the model.
+
+        All features must be prepended with `%` to be recognized by FreqAI internals.
+
+        :param df: strategy dataframe which will receive the features
+        :param period: period of the indicator - usage example:
+        dataframe["%-ema-period"] = ta.EMA(dataframe, timeperiod=period)
+        """
+
+        dataframe["%-rsi-period"] = ta.RSI(dataframe, timeperiod=period)
+        dataframe["%-mfi-period"] = ta.MFI(dataframe, timeperiod=period)
+        dataframe["%-adx-period"] = ta.ADX(dataframe, timeperiod=period)
+        dataframe["%-sma-period"] = ta.SMA(dataframe, timeperiod=period)
+        dataframe["%-ema-period"] = ta.EMA(dataframe, timeperiod=period)
+
+        return dataframe
+
+    def feature_engineering_expand_basic(self, dataframe: DataFrame, **kwargs) -> DataFrame:
+        """
+        *Only functional with FreqAI enabled strategies*
+        This function will automatically expand the defined features on the config defined
+        `include_timeframes`, `include_shifted_candles`, and `include_corr_pairs`.
+        In other words, a single feature defined in this function
+        will automatically expand to a total of
+        `include_timeframes` * `include_shifted_candles` * `include_corr_pairs`
+        numbers of features added to the model.
+
+        Features defined here will *not* be automatically duplicated on user defined
+        `indicator_periods_candles`
+
+        All features must be prepended with `%` to be recognized by FreqAI internals.
+
+        :param df: strategy dataframe which will receive the features
+        dataframe["%-pct-change"] = dataframe["close"].pct_change()
+        dataframe["%-ema-200"] = ta.EMA(dataframe, timeperiod=200)
+        """
+        dataframe["%-pct-change"] = dataframe["close"].pct_change()
+        dataframe["%-raw_volume"] = dataframe["volume"]
+        dataframe["%-raw_price"] = dataframe["close"]
+        return dataframe
+
+    def feature_engineering_standard(self, dataframe: DataFrame, **kwargs) -> DataFrame:
+        """
+        *Only functional with FreqAI enabled strategies*
+        This optional function will be called once with the dataframe of the base timeframe.
+        This is the final function to be called, which means that the dataframe entering this
+        function will contain all the features and columns created by all other
+        freqai_feature_engineering_* functions.
+
+        This function is a good place to do custom exotic feature extractions (e.g. tsfresh).
+        This function is a good place for any feature that should not be auto-expanded upon
+        (e.g. day of the week).
+
+        All features must be prepended with `%` to be recognized by FreqAI internals.
+
+        :param df: strategy dataframe which will receive the features
+        usage example: dataframe["%-day_of_week"] = (dataframe["date"].dt.dayofweek + 1) / 7
+        """
+        dataframe["%-day_of_week"] = (dataframe["date"].dt.dayofweek + 1) / 7
+        dataframe["%-hour_of_day"] = (dataframe["date"].dt.hour + 1) / 25
+        return dataframe
+
+    def set_freqai_targets(self, dataframe: DataFrame, **kwargs) -> DataFrame:
+        """
+        *Only functional with FreqAI enabled strategies*
+        Required function to set the targets for the model.
+        All targets must be prepended with `&` to be recognized by the FreqAI internals.
+
+        :param df: strategy dataframe which will receive the targets
+        usage example: dataframe["&-target"] = dataframe["close"].shift(-1) / dataframe["close"]
+        """
+        dataframe["&-s_close"] = (
+            dataframe["close"]
+            .shift(-self.freqai_info["feature_parameters"]["label_period_candles"])
+            .rolling(self.freqai_info["feature_parameters"]["label_period_candles"])
+            .mean()
+            / dataframe["close"]
+            - 1
+            )
+        return dataframe
+```
+
+Notice how the `feature_engineering_*()` is where [features](freqai-feature-engineering.md#feature-engineering) are added. Meanwhile `set_freqai_targets()` adds the labels/targets. A full example strategy is available in `templates/FreqaiExampleStrategy.py`.
+
+!!! Note
+    The `self.freqai.start()` function cannot be called outside the `populate_indicators()`.
+
+!!! Note
+    Features **must** be defined in `feature_engineering_*()`. Defining FreqAI features in `populate_indicators()`
+    will cause the algorithm to fail in live/dry mode. In order to add generalized features that are not associated with a specific pair or timeframe, you should use `feature_engineering_standard()`
+    (as exemplified in `freqtrade/templates/FreqaiExampleStrategy.py`).
+
+## Important dataframe key patterns
+
+Below are the values you can expect to include/use inside a typical strategy dataframe (`df[]`):
+
+|  DataFrame Key | Description |
+|------------|-------------|
+| `df['&*']` | Any dataframe column prepended with `&` in `set_freqai_targets()` is treated as a training target (label) inside FreqAI (typically following the naming convention `&-s*`). For example, to predict the close price 40 candles into the future, you would set `df['&-s_close'] = df['close'].shift(-self.freqai_info["feature_parameters"]["label_period_candles"])` with `"label_period_candles": 40` in the config. FreqAI makes the predictions and gives them back under the same key (`df['&-s_close']`) to be used in `populate_entry/exit_trend()`. <br> **Datatype:** Depends on the output of the model.
+| `df['&*_std/mean']` | Standard deviation and mean values of the defined labels during training (or live tracking with `fit_live_predictions_candles`). Commonly used to understand the rarity of a prediction (use the z-score as shown in `templates/FreqaiExampleStrategy.py` and explained [here](#creating-a-dynamic-target-threshold) to evaluate how often a particular prediction was observed during training or historically with `fit_live_predictions_candles`). <br> **Datatype:** Float.
+| `df['do_predict']` | Indication of an outlier data point. The return value is integer between -2 and 2, which lets you know if the prediction is trustworthy or not. `do_predict==1` means that the prediction is trustworthy. If the Dissimilarity Index (DI, see details [here](freqai-feature-engineering.md#identifying-outliers-with-the-dissimilarity-index-di)) of the input data point is above the threshold defined in the config, FreqAI will subtract 1 from `do_predict`, resulting in `do_predict==0`. If `use_SVM_to_remove_outliers` is active, the Support Vector Machine (SVM, see details [here](freqai-feature-engineering.md#identifying-outliers-using-a-support-vector-machine-svm)) may also detect outliers in training and prediction data. In this case, the SVM will also subtract 1 from `do_predict`. If the input data point was considered an outlier by the SVM but not by the DI, or vice versa, the result will be `do_predict==0`. If both the DI and the SVM considers the input data point to be an outlier, the result will be `do_predict==-1`. As with the SVM, if `use_DBSCAN_to_remove_outliers` is active, DBSCAN (see details [here](freqai-feature-engineering.md#identifying-outliers-with-dbscan)) may also detect outliers and subtract 1 from `do_predict`. Hence, if both the SVM and DBSCAN are active and identify a datapoint that was above the DI threshold as an outlier, the result will be `do_predict==-2`. A particular case is when `do_predict == 2`, which means that the model has expired due to exceeding `expired_hours`. <br> **Datatype:** Integer between -2 and 2.
+| `df['DI_values']` | Dissimilarity Index (DI) values are proxies for the level of confidence FreqAI has in the prediction. A lower DI means the prediction is close to the training data, i.e., higher prediction confidence. See details about the DI [here](freqai-feature-engineering.md#identifying-outliers-with-the-dissimilarity-index-di). <br> **Datatype:** Float.
+| `df['%*']` | Any dataframe column prepended with `%` in `feature_engineering_*()` is treated as a training feature. For example, you can include the RSI in the training feature set (similar to in `templates/FreqaiExampleStrategy.py`) by setting `df['%-rsi']`. See more details on how this is done [here](freqai-feature-engineering.md). <br> **Note:** Since the number of features prepended with `%` can multiply very quickly (10s of thousands of features are easily engineered using the multiplictative functionality of, e.g., `include_shifted_candles` and `include_timeframes` as described in the [parameter table](freqai-parameter-table.md)), these features are removed from the dataframe that is returned from FreqAI to the strategy. To keep a particular type of feature for plotting purposes, you would prepend it with `%%` (see details below). <br> **Datatype:** Depends on the feature created by the user.
+| `df['%%*']` | Any dataframe column prepended with `%%` in `feature_engineering_*()` is treated as a training feature, just the same as the above `%` prepend. However, in this case, the features are returned back to the strategy for FreqUI/plot-dataframe plotting and monitoring in Dry/Live/Backtesting <br> **Datatype:** Depends on the feature created by the user. Please note that features created in `feature_engineering_expand()` will have automatic FreqAI naming schemas depending on the expansions that you configured (i.e. `include_timeframes`, `include_corr_pairlist`, `indicators_periods_candles`, `include_shifted_candles`). So if you want to plot `%%-rsi` from `feature_engineering_expand_all()`, the final naming scheme for your plotting config would be: `%%-rsi-period_10_ETH/USDT:USDT_1h` for the `rsi` feature with `period=10`, `timeframe=1h`, and `pair=ETH/USDT:USDT` (the `:USDT` is added if you are using futures pairs). It is useful to simply add `print(dataframe.columns)` in your `populate_indicators()` after `self.freqai.start()` to see the full list of available features that are returned to the strategy for plotting purposes.
+
+## Setting the `startup_candle_count`
+
+The `startup_candle_count` in the FreqAI strategy needs to be set up in the same way as in the standard Freqtrade strategy (see details [here](strategy-customization.md#strategy-startup-period)). This value is used by Freqtrade to ensure that a sufficient amount of data is provided when calling the `dataprovider`, to avoid any NaNs at the beginning of the first training. You can easily set this value by identifying the longest period (in candle units) which is passed to the indicator creation functions (e.g., TA-Lib functions). In the presented example, `startup_candle_count` is 20 since this is the maximum value in `indicators_periods_candles`.
+
+!!! Note
+    There are instances where the TA-Lib functions actually require more data than just the passed `period` or else the feature dataset gets populated with NaNs. Anecdotally, multiplying the `startup_candle_count` by 2 always leads to a fully NaN free training dataset. Hence, it is typically safest to multiply the expected `startup_candle_count` by 2. Look out for this log message to confirm that the data is clean:
+
+    ```
+    2022-08-31 15:14:04 - freqtrade.freqai.data_kitchen - INFO - dropped 0 training points due to NaNs in populated dataset 4319.
+    ```
+
+## Creating a dynamic target threshold
+
+Deciding when to enter or exit a trade can be done in a dynamic way to reflect current market conditions. FreqAI allows you to return additional information from the training of a model (more info [here](freqai-feature-engineering.md#returning-additional-info-from-training)). For example, the `&*_std/mean` return values describe the statistical distribution of the target/label *during the most recent training*. Comparing a given prediction to these values allows you to know the rarity of the prediction. In `templates/FreqaiExampleStrategy.py`, the `target_roi` and  `sell_roi` are defined to be 1.25 z-scores away from the mean which causes predictions that are closer to the mean to be filtered out.
+
+```python
+dataframe["target_roi"] = dataframe["&-s_close_mean"] + dataframe["&-s_close_std"] * 1.25
+dataframe["sell_roi"] = dataframe["&-s_close_mean"] - dataframe["&-s_close_std"] * 1.25
+```
+
+To consider the population of *historical predictions* for creating the dynamic target instead of information from the training as discussed above, you would set `fit_live_predictions_candles` in the config to the number of historical prediction candles you wish to use to generate target statistics.
+
+```json
+    "freqai": {
+        "fit_live_predictions_candles": 300,
+    }
+```
+
+If this value is set, FreqAI will initially use the predictions from the training data and subsequently begin introducing real prediction data as it is generated. FreqAI will save this historical data to be reloaded if you stop and restart a model with the same `identifier`.
+
+## Using different prediction models
+
+FreqAI has multiple example prediction model libraries that are ready to be used as is via the flag `--freqaimodel`. These libraries include `CatBoost`, `LightGBM`, and `XGBoost` regression, classification, and multi-target models, and can be found in `freqai/prediction_models/`.
+
+Regression and classification models differ in what targets they predict - a regression model will predict a target of continuous values, for example what price BTC will be at tomorrow, whilst a classifier will predict a target of discrete values, for example if the price of BTC will go up tomorrow or not. This means that you have to specify your targets differently depending on which model type you are using (see details [below](#setting-model-targets)).
+
+All of the aforementioned model libraries implement gradient boosted decision tree algorithms. They all work on the principle of ensemble learning, where predictions from multiple simple learners are combined to get a final prediction that is more stable and generalized. The simple learners in this case are decision trees. Gradient boosting refers to the method of learning, where each simple learner is built in sequence - the subsequent learner is used to improve on the error from the previous learner. If you want to learn more about the different model libraries you can find the information in their respective docs:
+
+* CatBoost: https://catboost.ai/en/docs/
+* LightGBM: https://lightgbm.readthedocs.io/en/v3.3.2/#
+* XGBoost: https://xgboost.readthedocs.io/en/stable/#
+
+There are also numerous online articles describing and comparing the algorithms. Some relatively lightweight examples would be [CatBoost vs. LightGBM vs. XGBoost — Which is the best algorithm?](https://towardsdatascience.com/catboost-vs-lightgbm-vs-xgboost-c80f40662924#:~:text=In%20CatBoost%2C%20symmetric%20trees%2C%20or,the%20same%20depth%20can%20differ.) and [XGBoost, LightGBM or CatBoost — which boosting algorithm should I use?](https://medium.com/riskified-technology/xgboost-lightgbm-or-catboost-which-boosting-algorithm-should-i-use-e7fda7bb36bc). Keep in mind that the performance of each model is highly dependent on the application and so any reported metrics might not be true for your particular use of the model.
+
+Apart from the models already available in FreqAI, it is also possible to customize and create your own prediction models using the `IFreqaiModel` class. You are encouraged to inherit `fit()`, `train()`, and `predict()` to customize various aspects of the training procedures. You can place custom FreqAI models in `user_data/freqaimodels` - and freqtrade will pick them up from there based on the provided `--freqaimodel` name - which has to correspond to the class name of your custom model.
+Make sure to use unique names to avoid overriding built-in models.
+
+### Setting model targets
+
+#### Regressors
+
+If you are using a regressor, you need to specify a target that has continuous values. FreqAI includes a variety of regressors, such as the `CatboostRegressor`via the flag `--freqaimodel CatboostRegressor`. An example of how you could set a regression target for predicting the price 100 candles into the future would be
+
+```python
+df['&s-close_price'] = df['close'].shift(-100)
+```
+
+If you want to predict multiple targets, you need to define multiple labels using the same syntax as shown above.
+
+#### Classifiers
+
+If you are using a classifier, you need to specify a target that has discrete values. FreqAI includes a variety of classifiers, such as the `CatboostClassifier` via the flag `--freqaimodel CatboostClassifier`. If you elects to use a classifier, the classes need to be set using strings. For example, if you want to predict if the price 100 candles into the future goes up or down you would set
+
+```python
+df['&s-up_or_down'] = np.where( df["close"].shift(-100) > df["close"], 'up', 'down')
+```
+
+If you want to predict multiple targets you must specify all labels in the same label column. You could, for example, add the label `same` to define where the price was unchanged by setting
+
+```python
+df['&s-up_or_down'] = np.where( df["close"].shift(-100) > df["close"], 'up', 'down')
+df['&s-up_or_down'] = np.where( df["close"].shift(-100) == df["close"], 'same', df['&s-up_or_down'])
+```
+
+## PyTorch Module
+
+### Quick start
+
+The easiest way to quickly run a pytorch model is with the following command (for regression task):
+
+```bash
+freqtrade trade --config config_examples/config_freqai.example.json --strategy FreqaiExampleStrategy --freqaimodel PyTorchMLPRegressor --strategy-path freqtrade/templates 
+```
+
+!!! Note "Installation/docker"
+    The PyTorch module requires large packages such as `torch`, which should be explicitly requested during `./setup.sh -i` by answering "y" to the question "Do you also want dependencies for freqai-rl or PyTorch (~700mb additional space required) [y/N]?".
+    Users who prefer docker should ensure they use the docker image appended with `_freqaitorch`.
+    We do provide an explicit docker-compose file for this in `docker/docker-compose-freqai.yml` - which can be used via `docker compose -f docker/docker-compose-freqai.yml run ...` - or can be copied to replace the original docker file.
+    This docker-compose file also contains a (disabled) section to enable GPU resources within docker containers. This obviously assumes the system has GPU resources available.
+
+    PyTorch dropped support for macOS x64 (intel based Apple devices) in version 2.3. Subsequently, freqtrade also dropped support for PyTorch on this platform.
+
+### Structure
+
+#### Model
+
+You can construct your own Neural Network architecture in PyTorch by simply defining your `nn.Module` class inside your custom [`IFreqaiModel` file](#using-different-prediction-models) and then using that class in your `def train()` function. Here is an example of logistic regression model implementation using PyTorch (should be used with nn.BCELoss criterion) for classification tasks.
+
+```python
+
+class LogisticRegression(nn.Module):
+    def __init__(self, input_size: int):
+        super().__init__()
+        # Define your layers
+        self.linear = nn.Linear(input_size, 1)
+        self.activation = nn.Sigmoid()
+
+    def forward(self, x: torch.Tensor) -> torch.Tensor:
+        # Define the forward pass
+        out = self.linear(x)
+        out = self.activation(out)
+        return out
+
+class MyCoolPyTorchClassifier(BasePyTorchClassifier):
+    """
+    This is a custom IFreqaiModel showing how a user might setup their own 
+    custom Neural Network architecture for their training.
+    """
+
+    @property
+    def data_convertor(self) -> PyTorchDataConvertor:
+        return DefaultPyTorchDataConvertor(target_tensor_type=torch.float)
+
+    def __init__(self, **kwargs) -> None:
+        super().__init__(**kwargs)
+        config = self.freqai_info.get("model_training_parameters", {})
+        self.learning_rate: float = config.get("learning_rate",  3e-4)
+        self.model_kwargs: dict[str, Any] = config.get("model_kwargs",  {})
+        self.trainer_kwargs: dict[str, Any] = config.get("trainer_kwargs",  {})
+
+    def fit(self, data_dictionary: dict, dk: FreqaiDataKitchen, **kwargs) -> Any:
+        """
+        User sets up the training and test data to fit their desired model here
+        :param data_dictionary: the dictionary holding all data for train, test,
+            labels, weights
+        :param dk: The datakitchen object for the current coin/model
+        """
+
+        class_names = self.get_class_names()
+        self.convert_label_column_to_int(data_dictionary, dk, class_names)
+        n_features = data_dictionary["train_features"].shape[-1]
+        model = LogisticRegression(
+            input_dim=n_features
+        )
+        model.to(self.device)
+        optimizer = torch.optim.AdamW(model.parameters(), lr=self.learning_rate)
+        criterion = torch.nn.CrossEntropyLoss()
+        init_model = self.get_init_model(dk.pair)
+        trainer = PyTorchModelTrainer(
+            model=model,
+            optimizer=optimizer,
+            criterion=criterion,
+            model_meta_data={"class_names": class_names},
+            device=self.device,
+            init_model=init_model,
+            data_convertor=self.data_convertor,
+            **self.trainer_kwargs,
+        )
+        trainer.fit(data_dictionary, self.splits)
+        return trainer
+
+```
+
+#### Trainer
+
+The `PyTorchModelTrainer` performs the idiomatic PyTorch train loop:
+Define our model, loss function, and optimizer, and then move them to the appropriate device (GPU or CPU). Inside the loop, we iterate through the batches in the dataloader, move the data to the device, compute the prediction and loss, backpropagate, and update the model parameters using the optimizer. 
+
+In addition, the trainer is responsible for the following:
+ - saving and loading the model
+ - converting the data from `pandas.DataFrame` to `torch.Tensor`. 
+
+#### Integration with Freqai module 
+
+Like all freqai models, PyTorch models inherit `IFreqaiModel`. `IFreqaiModel` declares three abstract methods: `train`, `fit`, and `predict`. we implement these methods in three levels of hierarchy.
+From top to bottom:
+
+1. `BasePyTorchModel` - Implements the `train` method. all `BasePyTorch*` inherit it. responsible for general data preparation (e.g., data normalization) and calling the `fit` method. Sets `device` attribute used by children classes. Sets `model_type` attribute used by the parent class.
+2. `BasePyTorch*` -  Implements the `predict` method. Here, the `*` represents a group of algorithms, such as classifiers or regressors. responsible for data preprocessing, predicting, and postprocessing if needed.
+3. `PyTorch*Classifier` / `PyTorch*Regressor` - implements the `fit` method. responsible for the main train flaw, where we initialize the trainer and model objects.
+
+![image](assets/freqai_pytorch-diagram.png)
+
+#### Full example
+
+Building a PyTorch regressor using MLP (multilayer perceptron) model, MSELoss criterion, and AdamW optimizer.
+
+```python
+class PyTorchMLPRegressor(BasePyTorchRegressor):
+    def __init__(self, **kwargs) -> None:
+        super().__init__(**kwargs)
+        config = self.freqai_info.get("model_training_parameters", {})
+        self.learning_rate: float = config.get("learning_rate",  3e-4)
+        self.model_kwargs: dict[str, Any] = config.get("model_kwargs",  {})
+        self.trainer_kwargs: dict[str, Any] = config.get("trainer_kwargs",  {})
+
+    def fit(self, data_dictionary: dict, dk: FreqaiDataKitchen, **kwargs) -> Any:
+        n_features = data_dictionary["train_features"].shape[-1]
+        model = PyTorchMLPModel(
+            input_dim=n_features,
+            output_dim=1,
+            **self.model_kwargs
+        )
+        model.to(self.device)
+        optimizer = torch.optim.AdamW(model.parameters(), lr=self.learning_rate)
+        criterion = torch.nn.MSELoss()
+        init_model = self.get_init_model(dk.pair)
+        trainer = PyTorchModelTrainer(
+            model=model,
+            optimizer=optimizer,
+            criterion=criterion,
+            device=self.device,
+            init_model=init_model,
+            target_tensor_type=torch.float,
+            **self.trainer_kwargs,
+        )
+        trainer.fit(data_dictionary)
+        return trainer
+```
+
+Here we create a `PyTorchMLPRegressor` class that implements the `fit` method. The `fit` method specifies the training building blocks: model, optimizer, criterion, and trainer. We inherit both `BasePyTorchRegressor` and `BasePyTorchModel`, where the former implements the `predict` method that is suitable for our regression task, and the latter implements the train method.
+
+??? Note "Setting Class Names for Classifiers"
+    When using classifiers, the user must declare the class names (or targets) by overriding the `IFreqaiModel.class_names` attribute. This is achieved by setting `self.freqai.class_names` in the FreqAI strategy inside the `set_freqai_targets` method.
+    
+    For example, if you are using a binary classifier to predict price movements as up or down, you can set the class names as follows:
+    ```python
+    def set_freqai_targets(self, dataframe: DataFrame, metadata: dict, **kwargs) -> DataFrame:
+        self.freqai.class_names = ["down", "up"]
+        dataframe['&s-up_or_down'] = np.where(dataframe["close"].shift(-100) >
+                                                  dataframe["close"], 'up', 'down')
+    
+        return dataframe
+    ```
+    To see a full example, you can refer to the [classifier test strategy class](https://github.com/freqtrade/freqtrade/blob/develop/tests/strategy/strats/freqai_test_classifier.py).
+
+
+#### Improving performance with `torch.compile()`
+
+Torch provides a `torch.compile()` method that can be used to improve performance for specific GPU hardware. More details can be found [here](https://pytorch.org/tutorials/intermediate/torch_compile_tutorial.html). In brief, you simply wrap your `model` in `torch.compile()`:
+
+
+```python
+        model = PyTorchMLPModel(
+            input_dim=n_features,
+            output_dim=1,
+            **self.model_kwargs
+        )
+        model.to(self.device)
+        model = torch.compile(model)
+```
+
+Then proceed to use the model as normal. Keep in mind that doing this will remove eager execution, which means errors and tracebacks will not be informative.

md/freqai-developers.md

@@ -0,0 +1,78 @@
+# Development
+
+## Project architecture
+
+The architecture and functions of FreqAI are generalized to encourages development of unique features, functions, models, etc.
+
+The class structure and a detailed algorithmic overview is depicted in the following diagram:
+
+![image](assets/freqai_algorithm-diagram.jpg)
+
+As shown, there are three distinct objects comprising FreqAI:
+
+* **IFreqaiModel** - A singular persistent object containing all the necessary logic to collect, store, and process data, engineer features, run training, and inference models.
+* **FreqaiDataKitchen** - A non-persistent object which is created uniquely for each unique asset/model. Beyond metadata, it also contains a variety of data processing tools.
+* **FreqaiDataDrawer** - A singular persistent object containing all the historical predictions, models, and save/load methods.
+
+There are a variety of built-in [prediction models](freqai-configuration.md#using-different-prediction-models) which inherit directly from `IFreqaiModel`. Each of these models have full access to all methods in `IFreqaiModel` and can therefore override any of those functions at will. However, advanced users will likely stick to overriding `fit()`, `train()`, `predict()`, and `data_cleaning_train/predict()`.
+
+## Data handling
+
+FreqAI aims to organize model files, prediction data, and meta data in a way that simplifies post-processing and enhances crash resilience by automatic data reloading. The data is saved in a file structure,`user_data_dir/models/`, which contains all the data associated with the trainings and backtests. The `FreqaiDataKitchen()` relies heavily on the file structure for proper training and inferencing and should therefore not be manually modified.
+
+### File structure
+
+The file structure is automatically generated based on the model `identifier` set in the [config](freqai-configuration.md#setting-up-the-configuration-file). The following structure shows where the data is stored for post processing:
+
+| Structure | Description |
+|-----------|-------------|
+| `config_*.json` | A copy of the model specific configuration file. |
+| `historic_predictions.pkl` | A file containing all historic predictions generated during the lifetime of the `identifier` model during live deployment. `historic_predictions.pkl` is used to reload the model after a crash or a config change. A backup file is always held in case of corruption on the main file. FreqAI **automatically** detects corruption and replaces the corrupted file with the backup. |
+| `pair_dictionary.json` | A file containing the training queue as well as the on disk location of the most recently trained model. |
+| `sub-train-*_TIMESTAMP` | A folder containing all the files associated with a single model, such as: <br>
+|| `*_metadata.json` - Metadata for the model, such as normalization max/min, expected training feature list, etc. <br>
+|| `*_model.*` - The model file saved to disk for reloading from a crash. Can be `joblib` (typical boosting libs), `zip` (stable_baselines), `hd5` (keras type), etc. <br>
+|| `*_pca_object.pkl` - The [Principal component analysis (PCA)](freqai-feature-engineering.md#data-dimensionality-reduction-with-principal-component-analysis) transform (if `principal_component_analysis: True` is set in the config) which will be used to transform unseen prediction features. <br>
+|| `*_svm_model.pkl` - The [Support Vector Machine (SVM)](freqai-feature-engineering.md#identifying-outliers-using-a-support-vector-machine-svm) model (if `use_SVM_to_remove_outliers: True` is set in the config) which is used to detect outliers in unseen prediction features. <br>
+|| `*_trained_df.pkl` - The dataframe containing all the training features used to train the `identifier` model. This is used for computing the [Dissimilarity Index (DI)](freqai-feature-engineering.md#identifying-outliers-with-the-dissimilarity-index-di) and can also be used for post-processing. <br>
+|| `*_trained_dates.df.pkl` - The dates associated with the `trained_df.pkl`, which is useful for post-processing. |
+
+The example file structure would look like this:
+
+```
+├── models
+│   └── unique-id
+│       ├── config_freqai.example.json
+│       ├── historic_predictions.backup.pkl
+│       ├── historic_predictions.pkl
+│       ├── pair_dictionary.json
+│       ├── sub-train-1INCH_1662821319
+│       │   ├── cb_1inch_1662821319_metadata.json
+│       │   ├── cb_1inch_1662821319_model.joblib
+│       │   ├── cb_1inch_1662821319_pca_object.pkl
+│       │   ├── cb_1inch_1662821319_svm_model.joblib
+│       │   ├── cb_1inch_1662821319_trained_dates_df.pkl
+│       │   └── cb_1inch_1662821319_trained_df.pkl
+│       ├── sub-train-1INCH_1662821371
+│       │   ├── cb_1inch_1662821371_metadata.json
+│       │   ├── cb_1inch_1662821371_model.joblib
+│       │   ├── cb_1inch_1662821371_pca_object.pkl
+│       │   ├── cb_1inch_1662821371_svm_model.joblib
+│       │   ├── cb_1inch_1662821371_trained_dates_df.pkl
+│       │   └── cb_1inch_1662821371_trained_df.pkl
+│       ├── sub-train-ADA_1662821344
+│       │   ├── cb_ada_1662821344_metadata.json
+│       │   ├── cb_ada_1662821344_model.joblib
+│       │   ├── cb_ada_1662821344_pca_object.pkl
+│       │   ├── cb_ada_1662821344_svm_model.joblib
+│       │   ├── cb_ada_1662821344_trained_dates_df.pkl
+│       │   └── cb_ada_1662821344_trained_df.pkl
+│       └── sub-train-ADA_1662821399
+│           ├── cb_ada_1662821399_metadata.json
+│           ├── cb_ada_1662821399_model.joblib
+│           ├── cb_ada_1662821399_pca_object.pkl
+│           ├── cb_ada_1662821399_svm_model.joblib
+│           ├── cb_ada_1662821399_trained_dates_df.pkl
+│           └── cb_ada_1662821399_trained_df.pkl
+
+```

md/freqai-feature-engineering.md

@@ -0,0 +1,408 @@
+# Feature engineering
+
+## Defining the features
+
+Low level feature engineering is performed in the user strategy within a set of functions called `feature_engineering_*`. These function set the `base features` such as, `RSI`, `MFI`, `EMA`, `SMA`, time of day, volume, etc. The `base features` can be custom indicators or they can be imported from any technical-analysis library that you can find. FreqAI is equipped with a set of functions to simplify rapid large-scale feature engineering:
+
+|  Function | Description |
+|---------------|-------------|
+| `feature_engineering_expand_all()` | This optional function will automatically expand the defined features on the config defined `indicator_periods_candles`, `include_timeframes`, `include_shifted_candles`, and `include_corr_pairs`.
+| `feature_engineering_expand_basic()` | This optional function will automatically expand the defined features on the config defined `include_timeframes`, `include_shifted_candles`, and `include_corr_pairs`. Note: this function does *not* expand across `indicator_periods_candles`.
+| `feature_engineering_standard()` | This optional function will be called once with the dataframe of the base timeframe. This is the final function to be called, which means that the dataframe entering this function will contain all the features and columns from the base asset created by the other `feature_engineering_expand` functions. This function is a good place to do custom exotic feature extractions (e.g. tsfresh). This function is also a good place for any feature that should not be auto-expanded upon (e.g., day of the week).
+| `set_freqai_targets()` | Required function to set the targets for the model. All targets must be prepended with `&` to be recognized by the FreqAI internals.
+
+Meanwhile, high level feature engineering is handled within `"feature_parameters":{}` in the FreqAI config. Within this file, it is possible to decide large scale feature expansions on top of the `base_features` such as "including correlated pairs" or "including informative timeframes" or even "including recent candles."
+
+It is advisable to start from the template `feature_engineering_*` functions in the source provided example strategy (found in `templates/FreqaiExampleStrategy.py`) to ensure that the feature definitions are following the correct conventions. Here is an example of how to set the indicators and labels in the strategy:
+
+```python
+    def feature_engineering_expand_all(self, dataframe: DataFrame, period, metadata, **kwargs) -> DataFrame:
+        """
+        *Only functional with FreqAI enabled strategies*
+        This function will automatically expand the defined features on the config defined
+        `indicator_periods_candles`, `include_timeframes`, `include_shifted_candles`, and
+        `include_corr_pairs`. In other words, a single feature defined in this function
+        will automatically expand to a total of
+        `indicator_periods_candles` * `include_timeframes` * `include_shifted_candles` *
+        `include_corr_pairs` numbers of features added to the model.
+
+        All features must be prepended with `%` to be recognized by FreqAI internals.
+
+        Access metadata such as the current pair/timeframe/period with:
+
+        `metadata["pair"]` `metadata["tf"]`  `metadata["period"]`
+
+        :param df: strategy dataframe which will receive the features
+        :param period: period of the indicator - usage example:
+        :param metadata: metadata of current pair
+        dataframe["%-ema-period"] = ta.EMA(dataframe, timeperiod=period)
+        """
+
+        dataframe["%-rsi-period"] = ta.RSI(dataframe, timeperiod=period)
+        dataframe["%-mfi-period"] = ta.MFI(dataframe, timeperiod=period)
+        dataframe["%-adx-period"] = ta.ADX(dataframe, timeperiod=period)
+        dataframe["%-sma-period"] = ta.SMA(dataframe, timeperiod=period)
+        dataframe["%-ema-period"] = ta.EMA(dataframe, timeperiod=period)
+
+        bollinger = qtpylib.bollinger_bands(
+            qtpylib.typical_price(dataframe), window=period, stds=2.2
+        )
+        dataframe["bb_lowerband-period"] = bollinger["lower"]
+        dataframe["bb_middleband-period"] = bollinger["mid"]
+        dataframe["bb_upperband-period"] = bollinger["upper"]
+
+        dataframe["%-bb_width-period"] = (
+            dataframe["bb_upperband-period"]
+            - dataframe["bb_lowerband-period"]
+        ) / dataframe["bb_middleband-period"]
+        dataframe["%-close-bb_lower-period"] = (
+            dataframe["close"] / dataframe["bb_lowerband-period"]
+        )
+
+        dataframe["%-roc-period"] = ta.ROC(dataframe, timeperiod=period)
+
+        dataframe["%-relative_volume-period"] = (
+            dataframe["volume"] / dataframe["volume"].rolling(period).mean()
+        )
+
+        return dataframe
+
+    def feature_engineering_expand_basic(self, dataframe: DataFrame, metadata, **kwargs) -> DataFrame:
+        """
+        *Only functional with FreqAI enabled strategies*
+        This function will automatically expand the defined features on the config defined
+        `include_timeframes`, `include_shifted_candles`, and `include_corr_pairs`.
+        In other words, a single feature defined in this function
+        will automatically expand to a total of
+        `include_timeframes` * `include_shifted_candles` * `include_corr_pairs`
+        numbers of features added to the model.
+
+        Features defined here will *not* be automatically duplicated on user defined
+        `indicator_periods_candles`
+
+        Access metadata such as the current pair/timeframe with:
+
+        `metadata["pair"]` `metadata["tf"]`
+
+        All features must be prepended with `%` to be recognized by FreqAI internals.
+
+        :param df: strategy dataframe which will receive the features
+        :param metadata: metadata of current pair
+        dataframe["%-pct-change"] = dataframe["close"].pct_change()
+        dataframe["%-ema-200"] = ta.EMA(dataframe, timeperiod=200)
+        """
+        dataframe["%-pct-change"] = dataframe["close"].pct_change()
+        dataframe["%-raw_volume"] = dataframe["volume"]
+        dataframe["%-raw_price"] = dataframe["close"]
+        return dataframe
+
+    def feature_engineering_standard(self, dataframe: DataFrame, metadata, **kwargs) -> DataFrame:
+        """
+        *Only functional with FreqAI enabled strategies*
+        This optional function will be called once with the dataframe of the base timeframe.
+        This is the final function to be called, which means that the dataframe entering this
+        function will contain all the features and columns created by all other
+        freqai_feature_engineering_* functions.
+
+        This function is a good place to do custom exotic feature extractions (e.g. tsfresh).
+        This function is a good place for any feature that should not be auto-expanded upon
+        (e.g. day of the week).
+
+        Access metadata such as the current pair with:
+
+        `metadata["pair"]`
+
+        All features must be prepended with `%` to be recognized by FreqAI internals.
+
+        :param df: strategy dataframe which will receive the features
+        :param metadata: metadata of current pair
+        usage example: dataframe["%-day_of_week"] = (dataframe["date"].dt.dayofweek + 1) / 7
+        """
+        dataframe["%-day_of_week"] = (dataframe["date"].dt.dayofweek + 1) / 7
+        dataframe["%-hour_of_day"] = (dataframe["date"].dt.hour + 1) / 25
+        return dataframe
+
+    def set_freqai_targets(self, dataframe: DataFrame, metadata, **kwargs) -> DataFrame:
+        """
+        *Only functional with FreqAI enabled strategies*
+        Required function to set the targets for the model.
+        All targets must be prepended with `&` to be recognized by the FreqAI internals.
+
+        Access metadata such as the current pair with:
+
+        `metadata["pair"]`
+
+        :param df: strategy dataframe which will receive the targets
+        :param metadata: metadata of current pair
+        usage example: dataframe["&-target"] = dataframe["close"].shift(-1) / dataframe["close"]
+        """
+        dataframe["&-s_close"] = (
+            dataframe["close"]
+            .shift(-self.freqai_info["feature_parameters"]["label_period_candles"])
+            .rolling(self.freqai_info["feature_parameters"]["label_period_candles"])
+            .mean()
+            / dataframe["close"]
+            - 1
+            )
+        
+        return dataframe
+```
+
+In the presented example, the user does not wish to pass the `bb_lowerband` as a feature to the model,
+and has therefore not prepended it with `%`. The user does, however, wish to pass `bb_width` to the
+model for training/prediction and has therefore prepended it with `%`.
+
+After having defined the `base features`, the next step is to expand upon them using the powerful `feature_parameters` in the configuration file:
+
+```json
+    "freqai": {
+        //...
+        "feature_parameters" : {
+            "include_timeframes": ["5m","15m","4h"],
+            "include_corr_pairlist": [
+                "ETH/USD",
+                "LINK/USD",
+                "BNB/USD"
+            ],
+            "label_period_candles": 24,
+            "include_shifted_candles": 2,
+            "indicator_periods_candles": [10, 20]
+        },
+        //...
+    }
+```
+
+The `include_timeframes` in the config above are the timeframes (`tf`) of each call to `feature_engineering_expand_*()` in the strategy. In the presented case, the user is asking for the `5m`, `15m`, and `4h` timeframes of the `rsi`, `mfi`, `roc`, and `bb_width` to be included in the feature set.
+
+You can ask for each of the defined features to be included also for informative pairs using the `include_corr_pairlist`. This means that the feature set will include all the features from `feature_engineering_expand_*()` on all the `include_timeframes` for each of the correlated pairs defined in the config (`ETH/USD`, `LINK/USD`, and `BNB/USD` in the presented example).
+
+`include_shifted_candles` indicates the number of previous candles to include in the feature set. For example, `include_shifted_candles: 2` tells FreqAI to include the past 2 candles for each of the features in the feature set.
+
+In total, the number of features the user of the presented example strategy has created is: length of `include_timeframes` * no. features in `feature_engineering_expand_*()` * length of `include_corr_pairlist` * no. `include_shifted_candles` * length of `indicator_periods_candles`
+ $= 3 * 3 * 3 * 2 * 2 = 108$.
+ 
+ !!! note "Learn more about creative feature engineering"
+    Check out our [medium article](https://emergentmethods.medium.com/freqai-from-price-to-prediction-6fadac18b665) geared toward helping users learn how to creatively engineer features.
+
+### Gain finer control over `feature_engineering_*` functions with `metadata`
+
+All `feature_engineering_*` and `set_freqai_targets()` functions are passed a `metadata` dictionary which contains information about the `pair`, `tf` (timeframe), and `period` that FreqAI is automating for feature building. As such, a user can use `metadata` inside `feature_engineering_*` functions as criteria for blocking/reserving features for certain timeframes, periods, pairs etc.
+
+```python
+def feature_engineering_expand_all(self, dataframe: DataFrame, period, metadata, **kwargs) -> DataFrame:
+    if metadata["tf"] == "1h":
+        dataframe["%-roc-period"] = ta.ROC(dataframe, timeperiod=period)
+```
+
+This will block `ta.ROC()` from being added to any timeframes other than `"1h"`.
+
+### Returning additional info from training
+
+Important metrics can be returned to the strategy at the end of each model training by assigning them to `dk.data['extra_returns_per_train']['my_new_value'] = XYZ` inside the custom prediction model class. 
+
+FreqAI takes the `my_new_value` assigned in this dictionary and expands it to fit the dataframe that is returned to the strategy. You can then use the returned metrics in your strategy through `dataframe['my_new_value']`. An example of how return values can be used in FreqAI are the `&*_mean` and `&*_std` values that are used to [created a dynamic target threshold](freqai-configuration.md#creating-a-dynamic-target-threshold).
+
+Another example, where the user wants to use live metrics from the trade database, is shown below:
+
+```json
+    "freqai": {
+        "extra_returns_per_train": {"total_profit": 4}
+    }
+```
+
+You need to set the standard dictionary in the config so that FreqAI can return proper dataframe shapes. These values will likely be overridden by the prediction model, but in the case where the model has yet to set them, or needs a default initial value, the pre-set values are what will be returned.
+
+### Weighting features for temporal importance
+
+FreqAI allows you to set a `weight_factor` to weight recent data more strongly than past data via an exponential function:
+
+$$ W_i = \exp(\frac{-i}{\alpha*n}) $$
+
+where $W_i$ is the weight of data point $i$ in a total set of $n$ data points. Below is a figure showing the effect of different weight factors on the data points in a feature set.
+
+![weight-factor](assets/freqai_weight-factor.jpg)
+
+## Building the data pipeline
+
+By default, FreqAI builds a dynamic pipeline based on user configuration settings. The default settings are robust and designed to work with a variety of methods. These two steps are a `MinMaxScaler(-1,1)` and a `VarianceThreshold` which removes any column that has 0 variance. Users can activate other steps with more configuration parameters. For example if users add `use_SVM_to_remove_outliers: true` to the `freqai` config, then FreqAI will automatically add the [`SVMOutlierExtractor`](#identifying-outliers-using-a-support-vector-machine-svm) to the pipeline. Likewise, users can add `principal_component_analysis: true` to the `freqai` config to activate PCA. The [DissimilarityIndex](#identifying-outliers-with-the-dissimilarity-index-di) is activated with `DI_threshold: 1`. Finally, noise can also be added to the data with `noise_standard_deviation: 0.1`. Finally, users can add [DBSCAN](#identifying-outliers-with-dbscan) outlier removal with `use_DBSCAN_to_remove_outliers: true`.
+
+!!! note "More information available"
+    Please review the [parameter table](freqai-parameter-table.md) for more information on these parameters.
+
+
+### Customizing the pipeline
+
+Users are encouraged to customize the data pipeline to their needs by building their own data pipeline. This can be done by simply setting `dk.feature_pipeline` to their desired `Pipeline` object inside their `IFreqaiModel` `train()` function, or if they prefer not to touch the `train()` function, they can override `define_data_pipeline`/`define_label_pipeline` functions in their `IFreqaiModel`:
+
+!!! note "More information available"
+    FreqAI uses the [`DataSieve`](https://github.com/emergentmethods/datasieve) pipeline, which follows the SKlearn pipeline API, but adds, among other features, coherence between the X, y, and sample_weight vector point removals, feature removal, feature name following. 
+
+```python
+from datasieve.transforms import SKLearnWrapper, DissimilarityIndex
+from datasieve.pipeline import Pipeline
+from sklearn.preprocessing import QuantileTransformer, StandardScaler
+from freqai.base_models import BaseRegressionModel
+
+
+class MyFreqaiModel(BaseRegressionModel):
+    """
+    Some cool custom model
+    """
+    def fit(self, data_dictionary: Dict, dk: FreqaiDataKitchen, **kwargs) -> Any:
+        """
+        My custom fit function
+        """
+        model = cool_model.fit()
+        return model
+
+    def define_data_pipeline(self) -> Pipeline:
+        """
+        User defines their custom feature pipeline here (if they wish)
+        """
+        feature_pipeline = Pipeline([
+            ('qt', SKLearnWrapper(QuantileTransformer(output_distribution='normal'))),
+            ('di', ds.DissimilarityIndex(di_threshold=1))
+        ])
+
+        return feature_pipeline
+    
+    def define_label_pipeline(self) -> Pipeline:
+        """
+        User defines their custom label pipeline here (if they wish)
+        """
+        label_pipeline = Pipeline([
+            ('qt', SKLearnWrapper(StandardScaler())),
+        ])
+
+        return label_pipeline
+```
+
+Here, you are defining the exact pipeline that will be used for your feature set during training and prediction. You can use *most* SKLearn transformation steps by wrapping them in the `SKLearnWrapper` class as shown above. In addition, you can use any of the transformations available in the [`DataSieve` library](https://github.com/emergentmethods/datasieve). 
+
+You can easily add your own transformation by creating a class that inherits from the datasieve `BaseTransform` and implementing your `fit()`, `transform()` and `inverse_transform()` methods:
+
+```python
+from datasieve.transforms.base_transform import BaseTransform
+# import whatever else you need
+
+class MyCoolTransform(BaseTransform):
+    def __init__(self, **kwargs):
+        self.param1 = kwargs.get('param1', 1)
+
+    def fit(self, X, y=None, sample_weight=None, feature_list=None, **kwargs):
+        # do something with X, y, sample_weight, or/and feature_list
+        return X, y, sample_weight, feature_list
+
+    def transform(self, X, y=None, sample_weight=None,
+                  feature_list=None, outlier_check=False, **kwargs):
+        # do something with X, y, sample_weight, or/and feature_list
+        return X, y, sample_weight, feature_list
+
+    def inverse_transform(self, X, y=None, sample_weight=None, feature_list=None, **kwargs):
+        # do/dont do something with X, y, sample_weight, or/and feature_list
+        return X, y, sample_weight, feature_list
+```
+
+!!! note "Hint"
+    You can define this custom class in the same file as your `IFreqaiModel`.
+
+### Migrating a custom `IFreqaiModel` to the new Pipeline
+
+If you have created your own custom `IFreqaiModel` with a custom `train()`/`predict()` function, *and* you still rely on `data_cleaning_train/predict()`, then you will need to migrate to the new pipeline. If your model does *not* rely on `data_cleaning_train/predict()`, then you do not need to worry about this migration.
+
+More details about the migration can be found [here](strategy_migration.md#freqai---new-data-pipeline).
+
+## Outlier detection
+
+Equity and crypto markets suffer from a high level of non-patterned noise in the form of outlier data points. FreqAI implements a variety of methods to identify such outliers and hence mitigate risk.
+
+### Identifying outliers with the Dissimilarity Index (DI)
+
+The Dissimilarity Index (DI) aims to quantify the uncertainty associated with each prediction made by the model. 
+
+You can tell FreqAI to remove outlier data points from the training/test data sets using the DI by including the following statement in the config:
+
+```json
+    "freqai": {
+        "feature_parameters" : {
+            "DI_threshold": 1
+        }
+    }
+```
+
+Which will add `DissimilarityIndex` step to your `feature_pipeline` and set the threshold to 1. The DI allows predictions which are outliers (not existent in the model feature space) to be thrown out due to low levels of certainty. To do so, FreqAI measures the distance between each training data point (feature vector), $X_{a}$, and all other training data points:
+
+$$ d_{ab} = \sqrt{\sum_{j=1}^p(X_{a,j}-X_{b,j})^2} $$
+
+where $d_{ab}$ is the distance between the normalized points $a$ and $b$, and $p$ is the number of features, i.e., the length of the vector $X$. The characteristic distance, $\overline{d}$, for a set of training data points is simply the mean of the average distances:
+
+$$ \overline{d} = \sum_{a=1}^n(\sum_{b=1}^n(d_{ab}/n)/n) $$
+
+$\overline{d}$ quantifies the spread of the training data, which is compared to the distance between a new prediction feature vectors, $X_k$ and all the training data:
+
+$$ d_k = \arg \min d_{k,i} $$
+
+This enables the estimation of the Dissimilarity Index as:
+
+$$ DI_k = d_k/\overline{d} $$
+
+You can tweak the DI through the `DI_threshold` to increase or decrease the extrapolation of the trained model. A higher `DI_threshold` means that the DI is more lenient and allows predictions further away from the training data to be used whilst a lower `DI_threshold` has the opposite effect and hence discards more predictions.
+
+Below is a figure that describes the DI for a 3D data set.
+
+![DI](assets/freqai_DI.jpg)
+
+### Identifying outliers using a Support Vector Machine (SVM)
+
+You can tell FreqAI to remove outlier data points from the training/test data sets using a Support Vector Machine (SVM) by including the following statement in the config:
+
+```json
+    "freqai": {
+        "feature_parameters" : {
+            "use_SVM_to_remove_outliers": true
+        }
+    }
+```
+
+Which will add `SVMOutlierExtractor` step to your `feature_pipeline`. The SVM will be trained on the training data and any data point that the SVM deems to be beyond the feature space will be removed.
+
+You can elect to provide additional parameters for the SVM, such as `shuffle`, and `nu` via the `feature_parameters.svm_params` dictionary in the config.
+
+The parameter `shuffle` is by default set to `False` to ensure consistent results. If it is set to `True`, running the SVM multiple times on the same data set might result in different outcomes due to `max_iter` being to low for the algorithm to reach the demanded `tol`. Increasing `max_iter` solves this issue but causes the procedure to take longer time.
+
+The parameter `nu`, *very* broadly, is the amount of data points that should be considered outliers and should be between 0 and 1.
+
+### Identifying outliers with DBSCAN
+
+You can configure FreqAI to use DBSCAN to cluster and remove outliers from the training/test data set or incoming outliers from predictions, by activating `use_DBSCAN_to_remove_outliers` in the config:
+
+```json
+    "freqai": {
+        "feature_parameters" : {
+            "use_DBSCAN_to_remove_outliers": true
+        }
+    }
+```
+
+Which will add the `DataSieveDBSCAN` step to your `feature_pipeline`. This is an unsupervised machine learning algorithm that clusters data without needing to know how many clusters there should be.
+
+Given a number of data points $N$, and a distance $\varepsilon$, DBSCAN clusters the data set by setting all data points that have $N-1$ other data points within a distance of $\varepsilon$ as *core points*. A data point that is within a distance of $\varepsilon$ from a *core point* but that does not have $N-1$ other data points within a distance of $\varepsilon$ from itself is considered an *edge point*. A cluster is then the collection of *core points* and *edge points*. Data points that have no other data points at a distance $<\varepsilon$ are considered outliers. The figure below shows a cluster with $N = 3$.
+
+![dbscan](assets/freqai_dbscan.jpg)
+
+FreqAI uses `sklearn.cluster.DBSCAN` (details are available on scikit-learn's webpage [here](https://scikit-learn.org/stable/modules/generated/sklearn.cluster.DBSCAN.html) (external website)) with `min_samples` ($N$) taken as 1/4 of the no. of time points (candles) in the feature set. `eps` ($\varepsilon$) is computed automatically as the elbow point in the *k-distance graph* computed from the nearest neighbors in the pairwise distances of all data points in the feature set.
+
+
+### Data dimensionality reduction with Principal Component Analysis
+
+You can reduce the dimensionality of your features by activating the principal_component_analysis in the config:
+
+```json
+    "freqai": {
+        "feature_parameters" : {
+            "principal_component_analysis": true
+        }
+    }
+```
+
+This will perform PCA on the features and reduce their dimensionality so that the explained variance of the data set is >= 0.999. Reducing data dimensionality makes training the model faster and hence allows for more up-to-date models.

md/freqai-parameter-table.md

@@ -0,0 +1,118 @@
+# Parameter table
+
+The table below will list all configuration parameters available for FreqAI. Some of the parameters are exemplified in `config_examples/config_freqai.example.json`.
+
+Mandatory parameters are marked as **Required** and have to be set in one of the suggested ways.
+
+### General configuration parameters
+
+|  Parameter | Description |
+|------------|-------------|
+|  |  **General configuration parameters within the `config.freqai` tree**
+| `freqai` | **Required.** <br> The parent dictionary containing all the parameters for controlling FreqAI. <br> **Datatype:** Dictionary.
+| `train_period_days` | **Required.** <br> Number of days to use for the training data (width of the sliding window). <br> **Datatype:** Positive integer.
+| `backtest_period_days` | **Required.** <br> Number of days to inference from the trained model before sliding the `train_period_days` window defined above, and retraining the model during backtesting (more info [here](freqai-running.md#backtesting)). This can be fractional days, but beware that the provided `timerange` will be divided by this number to yield the number of trainings necessary to complete the backtest. <br> **Datatype:** Float.
+| `identifier` | **Required.** <br> A unique ID for the current model. If models are saved to disk, the `identifier` allows for reloading specific pre-trained models/data. <br> **Datatype:** String.
+| `live_retrain_hours` | Frequency of retraining during dry/live runs. <br> **Datatype:** Float > 0. <br> Default: `0` (models retrain as often as possible).
+| `expiration_hours` | Avoid making predictions if a model is more than `expiration_hours` old. <br> **Datatype:** Positive integer. <br> Default: `0` (models never expire).
+| `purge_old_models` | Number of models to keep on disk (not relevant to backtesting). Default is 2, which means that dry/live runs will keep the latest 2 models on disk. Setting to 0 keeps all models. This parameter also accepts a boolean to maintain backwards compatibility. <br> **Datatype:** Integer. <br> Default: `2`.
+| `save_backtest_models` | Save models to disk when running backtesting. Backtesting operates most efficiently by saving the prediction data and reusing them directly for subsequent runs (when you wish to tune entry/exit parameters). Saving backtesting models to disk also allows to use the same model files for starting a dry/live instance with the same model `identifier`. <br> **Datatype:** Boolean. <br> Default: `False` (no models are saved).
+| `fit_live_predictions_candles` | Number of historical candles to use for computing target (label) statistics from prediction data, instead of from the training dataset (more information can be found [here](freqai-configuration.md#creating-a-dynamic-target-threshold)). <br> **Datatype:** Positive integer.
+| `continual_learning` | Use the final state of the most recently trained model as starting point for the new model, allowing for incremental learning (more information can be found [here](freqai-running.md#continual-learning)). Beware that this is currently a naive approach to incremental learning, and it has a high probability of overfitting/getting stuck in local minima while the market moves away from your model. We have the connections here primarily for experimental purposes and so that it is ready for more mature approaches to continual learning in chaotic systems like the crypto market. <br> **Datatype:** Boolean. <br> Default: `False`.
+| `write_metrics_to_disk` | Collect train timings, inference timings and cpu usage in json file. <br> **Datatype:** Boolean. <br> Default: `False`
+| `data_kitchen_thread_count` | <br> Designate the number of threads you want to use for data processing (outlier methods, normalization, etc.). This has no impact on the number of threads used for training. If user does not set it (default), FreqAI will use max number of threads - 2 (leaving 1 physical core available for Freqtrade bot and FreqUI) <br> **Datatype:** Positive integer.
+| `activate_tensorboard` | <br> Indicate whether or not to activate tensorboard for the tensorboard enabled modules (currently Reinforcment Learning, XGBoost, Catboost, and PyTorch). Tensorboard needs Torch installed, which means you will need the torch/RL docker image or you need to answer "yes" to the install question about whether or not you wish to install Torch. <br> **Datatype:** Boolean. <br> Default: `True`.
+| `wait_for_training_iteration_on_reload` | <br> When using /reload or ctrl-c, wait for the current training iteration to finish before completing graceful shutdown. If set to `False`, FreqAI will break the current training iteration, allowing you to shutdown gracefully more quickly, but you will lose your current training iteration. <br> **Datatype:** Boolean. <br> Default: `True`.
+
+### Feature parameters
+
+|  Parameter | Description |
+|------------|-------------|
+|  |  **Feature parameters within the `freqai.feature_parameters` sub dictionary**
+| `feature_parameters` | A dictionary containing the parameters used to engineer the feature set. Details and examples are shown [here](freqai-feature-engineering.md). <br> **Datatype:** Dictionary.
+| `include_timeframes` | A list of timeframes that all indicators in `feature_engineering_expand_*()` will be created for. The list is added as features to the base indicators dataset. <br> **Datatype:** List of timeframes (strings).
+| `include_corr_pairlist` | A list of correlated coins that FreqAI will add as additional features to all `pair_whitelist` coins. All indicators set in `feature_engineering_expand_*()` during feature engineering (see details [here](freqai-feature-engineering.md)) will be created for each correlated coin. The correlated coins features are added to the base indicators dataset. <br> **Datatype:** List of assets (strings).
+| `label_period_candles` | Number of candles into the future that the labels are created for. This can be used in `set_freqai_targets()` (see `templates/FreqaiExampleStrategy.py` for detailed usage). This parameter is not necessarily required, you can create custom labels and choose whether to make use of this parameter or not. Please see `templates/FreqaiExampleStrategy.py` to see the example usage. <br> **Datatype:** Positive integer.
+| `include_shifted_candles` | Add features from previous candles to subsequent candles with the intent of adding historical information. If used, FreqAI will duplicate and shift all features from the `include_shifted_candles` previous candles so that the information is available for the subsequent candle. <br> **Datatype:** Positive integer.
+| `weight_factor` | Weight training data points according to their recency (see details [here](freqai-feature-engineering.md#weighting-features-for-temporal-importance)). <br> **Datatype:** Positive float (typically < 1).
+| `indicator_max_period_candles` | **No longer used (#7325)**. Replaced by `startup_candle_count` which is set in the [strategy](freqai-configuration.md#building-a-freqai-strategy). `startup_candle_count` is timeframe independent and defines the maximum *period* used in `feature_engineering_*()` for indicator creation. FreqAI uses this parameter together with the maximum timeframe in `include_time_frames` to calculate how many data points to download such that the first data point does not include a NaN. <br> **Datatype:** Positive integer.
+| `indicator_periods_candles` | Time periods to calculate indicators for. The indicators are added to the base indicator dataset. <br> **Datatype:** List of positive integers.
+| `principal_component_analysis` | Automatically reduce the dimensionality of the data set using Principal Component Analysis. See details about how it works [here](freqai-feature-engineering.md#data-dimensionality-reduction-with-principal-component-analysis) <br> **Datatype:** Boolean. <br> Default: `False`.
+| `plot_feature_importances` | Create a feature importance plot for each model for the top/bottom `plot_feature_importances` number of features. Plot is stored in `user_data/models/<identifier>/sub-train-<COIN>_<timestamp>.html`. <br> **Datatype:** Integer. <br> Default: `0`.
+| `DI_threshold` | Activates the use of the Dissimilarity Index for outlier detection when set to > 0. See details about how it works [here](freqai-feature-engineering.md#identifying-outliers-with-the-dissimilarity-index-di). <br> **Datatype:** Positive float (typically < 1).
+| `use_SVM_to_remove_outliers` | Train a support vector machine to detect and remove outliers from the training dataset, as well as from incoming data points. See details about how it works [here](freqai-feature-engineering.md#identifying-outliers-using-a-support-vector-machine-svm). <br> **Datatype:** Boolean.
+| `svm_params` | All parameters available in Sklearn's `SGDOneClassSVM()`. See details about some select parameters [here](freqai-feature-engineering.md#identifying-outliers-using-a-support-vector-machine-svm). <br> **Datatype:** Dictionary.
+| `use_DBSCAN_to_remove_outliers` | Cluster data using the DBSCAN algorithm to identify and remove outliers from training and prediction data. See details about how it works [here](freqai-feature-engineering.md#identifying-outliers-with-dbscan). <br> **Datatype:** Boolean. 
+| `noise_standard_deviation` | If set, FreqAI adds noise to the training features with the aim of preventing overfitting. FreqAI generates random deviates from a gaussian distribution with a standard deviation of `noise_standard_deviation` and adds them to all data points. `noise_standard_deviation` should be kept relative to the normalized space, i.e., between -1 and 1. In other words, since data in FreqAI is always normalized to be between -1 and 1, `noise_standard_deviation: 0.05` would result in 32% of the data being randomly increased/decreased by more than 2.5% (i.e., the percent of data falling within the first standard deviation). <br> **Datatype:** Integer. <br> Default: `0`.
+| `outlier_protection_percentage` | Enable to prevent outlier detection methods from discarding too much data. If more than `outlier_protection_percentage` % of points are detected as outliers by the SVM or DBSCAN, FreqAI will log a warning message and ignore outlier detection, i.e., the original dataset will be kept intact. If the outlier protection is triggered, no predictions will be made based on the training dataset. <br> **Datatype:** Float. <br> Default: `30`.
+| `reverse_train_test_order` | Split the feature dataset (see below) and use the latest data split for training and test on historical split of the data. This allows the model to be trained up to the most recent data point, while avoiding overfitting. However, you should be careful to understand the unorthodox nature of this parameter before employing it. <br> **Datatype:** Boolean. <br> Default: `False` (no reversal).
+| `shuffle_after_split` | Split the data into train and test sets, and then shuffle both sets individually. <br> **Datatype:** Boolean. <br> Default: `False`.
+| `buffer_train_data_candles` | Cut `buffer_train_data_candles` off the beginning and end of the training data *after* the indicators were populated. The main example use is when predicting maxima and minima, the argrelextrema function  cannot know the maxima/minima at the edges of the timerange. To improve model accuracy, it is best to compute argrelextrema on the full timerange and then use this function to cut off the edges (buffer) by the kernel. In another case, if the targets are set to a shifted price movement, this buffer is unnecessary because the shifted candles at the end of the timerange will be NaN and FreqAI will automatically cut those off of the training dataset.<br> **Datatype:** Integer. <br> Default: `0`.
+
+### Data split parameters
+
+|  Parameter | Description |
+|------------|-------------|
+|  |  **Data split parameters within the `freqai.data_split_parameters` sub dictionary**
+| `data_split_parameters` | Include any additional parameters available from scikit-learn `test_train_split()`, which are shown [here](https://scikit-learn.org/stable/modules/generated/sklearn.model_selection.train_test_split.html) (external website). <br> **Datatype:** Dictionary.
+| `test_size` | The fraction of data that should be used for testing instead of training. <br> **Datatype:** Positive float < 1.
+| `shuffle` | Shuffle the training data points during training. Typically, to not remove the chronological order of data in time-series forecasting, this is set to `False`. <br> **Datatype:** Boolean. <br> Default: `False`.
+
+### Model training parameters
+
+|  Parameter | Description |
+|------------|-------------|
+|  |  **Model training parameters within the `freqai.model_training_parameters` sub dictionary**
+| `model_training_parameters` | A flexible dictionary that includes all parameters available by the selected model library. For example, if you use `LightGBMRegressor`, this dictionary can contain any parameter available by the `LightGBMRegressor` [here](https://lightgbm.readthedocs.io/en/latest/pythonapi/lightgbm.LGBMRegressor.html) (external website). If you select a different model, this dictionary can contain any parameter from that model. A list of the currently available models can be found [here](freqai-configuration.md#using-different-prediction-models).  <br> **Datatype:** Dictionary.
+| `n_estimators` | The number of boosted trees to fit in the training of the model. <br> **Datatype:** Integer.
+| `learning_rate` | Boosting learning rate during training of the model. <br> **Datatype:** Float.
+| `n_jobs`, `thread_count`, `task_type` | Set the number of threads for parallel processing and the `task_type` (`gpu` or `cpu`). Different model libraries use different parameter names. <br> **Datatype:** Float.
+
+### Reinforcement Learning parameters
+
+|  Parameter | Description |
+|------------|-------------|
+|  |  **Reinforcement Learning Parameters within the `freqai.rl_config` sub dictionary**
+| `rl_config` | A dictionary containing the control parameters for a Reinforcement Learning model. <br> **Datatype:** Dictionary.
+| `train_cycles` | Training time steps will be set based on the `train_cycles * number of training data points. <br> **Datatype:** Integer.
+| `max_trade_duration_candles`| Guides the agent training to keep trades below desired length. Example usage shown in `prediction_models/ReinforcementLearner.py` within the customizable `calculate_reward()` function. <br> **Datatype:** int.
+| `model_type` | Model string from stable_baselines3 or SBcontrib. Available strings include: `'TRPO', 'ARS', 'RecurrentPPO', 'MaskablePPO', 'PPO', 'A2C', 'DQN'`. User should ensure that `model_training_parameters` match those available to the corresponding stable_baselines3 model by visiting their documentation. [PPO doc](https://stable-baselines3.readthedocs.io/en/master/modules/ppo.html) (external website) <br> **Datatype:** string.
+| `policy_type` | One of the available policy types from stable_baselines3 <br> **Datatype:** string.
+| `max_training_drawdown_pct` | The maximum drawdown that the agent is allowed to experience during training. <br> **Datatype:** float. <br> Default: 0.8
+| `cpu_count` | Number of threads/cpus to dedicate to the Reinforcement Learning training process (depending on if `ReinforcementLearning_multiproc` is selected or not). Recommended to leave this untouched, by default, this value is set to the total number of physical cores minus 1. <br> **Datatype:** int. 
+| `model_reward_parameters` | Parameters used inside the customizable `calculate_reward()` function in `ReinforcementLearner.py` <br> **Datatype:** int.
+| `add_state_info` | Tell FreqAI to include state information in the feature set for training and inferencing. The current state variables include trade duration, current profit, trade position. This is only available in dry/live runs, and is automatically switched to false for backtesting. <br> **Datatype:** bool. <br> Default: `False`.
+| `net_arch` | Network architecture which is well described in [`stable_baselines3` doc](https://stable-baselines3.readthedocs.io/en/master/guide/custom_policy.html#examples). In summary: `[<shared layers>, dict(vf=[<non-shared value network layers>], pi=[<non-shared policy network layers>])]`. By default this is set to `[128, 128]`, which defines 2 shared hidden layers with 128 units each.
+| `randomize_starting_position` | Randomize the starting point of each episode to avoid overfitting. <br> **Datatype:** bool. <br> Default: `False`.
+| `drop_ohlc_from_features` | Do not include the normalized ohlc data in the feature set passed to the agent during training (ohlc will still be used for driving the environment in all cases) <br> **Datatype:** Boolean. <br> **Default:** `False`
+| `progress_bar` | Display a progress bar with the current progress, elapsed time and estimated remaining time. <br> **Datatype:** Boolean. <br> Default: `False`.
+
+### PyTorch parameters
+
+#### general
+
+|  Parameter | Description |
+|------------|-------------|
+|  |  **Model training parameters within the `freqai.model_training_parameters` sub dictionary**
+| `learning_rate` | Learning rate to be passed to the optimizer. <br> **Datatype:** float. <br> Default: `3e-4`.
+| `model_kwargs` | Parameters to be passed to the model class. <br> **Datatype:** dict. <br> Default: `{}`.
+| `trainer_kwargs` | Parameters to be passed to the trainer class. <br> **Datatype:** dict. <br> Default: `{}`.
+
+#### trainer_kwargs
+
+| Parameter    | Description |
+|--------------|-------------|
+|              |  **Model training parameters within the `freqai.model_training_parameters.model_kwargs` sub dictionary**
+| `n_epochs`   | The `n_epochs` parameter is a crucial setting in the PyTorch training loop that determines the number of times the entire training dataset will be used to update the model's parameters. An epoch represents one full pass through the entire training dataset. Overrides `n_steps`. Either `n_epochs` or `n_steps` must be set. <br><br> **Datatype:** int. optional. <br> Default: `10`.
+| `n_steps`    | An alternative way of setting `n_epochs` -  the number of training iterations to run. Iteration here refer to the number of times we call `optimizer.step()`. Ignored if `n_epochs` is set. A simplified version of the function: <br><br> n_epochs = n_steps / (n_obs / batch_size) <br><br> The motivation here is that `n_steps` is easier to optimize and keep stable across different n_obs - the number of data points.  <br> <br> **Datatype:** int. optional. <br> Default: `None`.
+| `batch_size` | The size of the batches to use during training. <br><br> **Datatype:** int. <br> Default: `64`.
+
+
+### Additional parameters
+
+|  Parameter | Description |
+|------------|-------------|
+|  |  **Extraneous parameters**
+| `freqai.keras` | If the selected model makes use of Keras (typical for TensorFlow-based prediction models), this flag needs to be activated so that the model save/loading follows Keras standards. <br> **Datatype:** Boolean. <br> Default: `False`.
+| `freqai.conv_width` | The width of a neural network input tensor. This replaces the need for shifting candles (`include_shifted_candles`) by feeding in historical data points as the second dimension of the tensor. Technically, this parameter can also be used for regressors, but it only adds computational overhead and does not change the model training/prediction. <br> **Datatype:** Integer. <br> Default: `2`.
+| `freqai.reduce_df_footprint` | Recast all numeric columns to float32/int32, with the objective of reducing ram/disk usage and decreasing train/inference timing. This parameter is set in the main level of the Freqtrade configuration file (not inside FreqAI). <br> **Datatype:** Boolean. <br> Default: `False`.

md/freqai-reinforcement-learning.md

@@ -0,0 +1,280 @@
+# Reinforcement Learning
+
+!!! Note "Installation size"
+    Reinforcement learning dependencies include large packages such as `torch`, which should be explicitly requested during `./setup.sh -i` by answering "y" to the question "Do you also want dependencies for freqai-rl (~700mb additional space required) [y/N]?".
+    Users who prefer docker should ensure they use the docker image appended with `_freqairl`.
+
+## Background and terminology
+
+### What is RL and why does FreqAI need it?
+
+Reinforcement learning involves two important components, the *agent* and the training *environment*. During agent training, the agent moves through historical data candle by candle, always making 1 of a set of actions: Long entry, long exit, short entry, short exit, neutral). During this training process, the environment tracks the performance of these actions and rewards the agent according to a custom user made `calculate_reward()` (here we offer a default reward for users to build on if they wish [details here](#creating-a-custom-reward-function)). The reward is used to train weights in a neural network.
+
+A second important component of the FreqAI RL implementation is the use of *state* information. State information is fed into the network at each step, including current profit, current position, and current trade duration. These are used to train the agent in the training environment, and to reinforce the agent in dry/live (this functionality is not available in backtesting). *FreqAI + Freqtrade is a perfect match for this reinforcing mechanism since this information is readily available in live deployments.*
+
+Reinforcement learning is a natural progression for FreqAI, since it adds a new layer of adaptivity and market reactivity that Classifiers and Regressors cannot match. However, Classifiers and Regressors have strengths that RL does not have such as robust predictions. Improperly trained RL agents may find "cheats" and "tricks" to maximize reward without actually winning any trades. For this reason, RL is more complex and demands a higher level of understanding than typical Classifiers and Regressors.
+
+### The RL interface
+
+With the current framework, we aim to expose the training environment via the common "prediction model" file, which is a user inherited `BaseReinforcementLearner` object (e.g. `freqai/prediction_models/ReinforcementLearner`). Inside this user class, the RL environment is available and customized via `MyRLEnv` as [shown below](#creating-a-custom-reward-function).
+
+We envision the majority of users focusing their effort on creative design of the `calculate_reward()` function [details here](#creating-a-custom-reward-function), while leaving the rest of the environment untouched. Other users may not touch the environment at all, and they will only play with the configuration settings and the powerful feature engineering that already exists in FreqAI. Meanwhile, we enable advanced users to create their own model classes entirely.
+
+The framework is built on stable_baselines3 (torch) and OpenAI gym for the base environment class. But generally speaking, the model class is well isolated. Thus, the addition of competing libraries can be easily integrated into the existing framework. For the environment, it is inheriting from `gym.Env` which means that it is necessary to write an entirely new environment in order to switch to a different library.
+
+### Important considerations
+
+As explained above, the agent is "trained" in an artificial trading "environment". In our case, that environment may seem quite similar to a real Freqtrade backtesting environment, but it is *NOT*. In fact, the RL training environment is much more simplified. It does not incorporate any of the complicated strategy logic, such as callbacks like `custom_exit`, `custom_stoploss`, leverage controls, etc. The RL environment is instead a very "raw" representation of the true market, where the agent has free will to learn the policy (read: stoploss, take profit, etc.) which is enforced by the `calculate_reward()`. Thus, it is important to consider that the agent training environment is not identical to the real world.
+
+## Running Reinforcement Learning
+
+Setting up and running a Reinforcement Learning model is the same as running a Regressor or Classifier. The same two flags, `--freqaimodel` and `--strategy`, must be defined on the command line:
+
+```bash
+freqtrade trade --freqaimodel ReinforcementLearner --strategy MyRLStrategy --config config.json
+```
+
+where `ReinforcementLearner` will use the templated `ReinforcementLearner` from `freqai/prediction_models/ReinforcementLearner` (or a custom user defined one located in `user_data/freqaimodels`). The strategy, on the other hand, follows the same base [feature engineering](freqai-feature-engineering.md) with `feature_engineering_*` as a typical Regressor. The difference lies in the creation of the targets, Reinforcement Learning doesn't require them. However, FreqAI requires a default (neutral) value to be set in the action column:
+
+```python
+    def set_freqai_targets(self, dataframe, **kwargs) -> DataFrame:
+        """
+        *Only functional with FreqAI enabled strategies*
+        Required function to set the targets for the model.
+        All targets must be prepended with `&` to be recognized by the FreqAI internals.
+
+        More details about feature engineering available:
+
+        https://www.freqtrade.io/en/latest/freqai-feature-engineering
+
+        :param df: strategy dataframe which will receive the targets
+        usage example: dataframe["&-target"] = dataframe["close"].shift(-1) / dataframe["close"]
+        """
+        # For RL, there are no direct targets to set. This is filler (neutral)
+        # until the agent sends an action.
+        dataframe["&-action"] = 0
+        return dataframe
+```
+
+Most of the function remains the same as for typical Regressors, however, the function below shows how the strategy must pass the raw price data to the agent so that it has access to raw OHLCV in the training environment:
+
+```python
+    def feature_engineering_standard(self, dataframe: DataFrame, **kwargs) -> DataFrame:
+        # The following features are necessary for RL models
+        dataframe[f"%-raw_close"] = dataframe["close"]
+        dataframe[f"%-raw_open"] = dataframe["open"]
+        dataframe[f"%-raw_high"] = dataframe["high"]
+        dataframe[f"%-raw_low"] = dataframe["low"]
+    return dataframe
+```
+
+Finally, there is no explicit "label" to make - instead it is necessary to assign the `&-action` column which will contain the agent's actions when accessed in `populate_entry/exit_trends()`. In the present example, the neutral action to 0. This value should align with the environment used. FreqAI provides two environments, both use 0 as the neutral action.
+
+After users realize there are no labels to set, they will soon understand that the agent is making its "own" entry and exit decisions. This makes strategy construction rather simple. The entry and exit signals come from the agent in the form of an integer - which are used directly to decide entries and exits in the strategy:
+
+```python
+    def populate_entry_trend(self, df: DataFrame, metadata: dict) -> DataFrame:
+
+        enter_long_conditions = [df["do_predict"] == 1, df["&-action"] == 1]
+
+        if enter_long_conditions:
+            df.loc[
+                reduce(lambda x, y: x & y, enter_long_conditions), ["enter_long", "enter_tag"]
+            ] = (1, "long")
+
+        enter_short_conditions = [df["do_predict"] == 1, df["&-action"] == 3]
+
+        if enter_short_conditions:
+            df.loc[
+                reduce(lambda x, y: x & y, enter_short_conditions), ["enter_short", "enter_tag"]
+            ] = (1, "short")
+
+        return df
+
+    def populate_exit_trend(self, df: DataFrame, metadata: dict) -> DataFrame:
+        exit_long_conditions = [df["do_predict"] == 1, df["&-action"] == 2]
+        if exit_long_conditions:
+            df.loc[reduce(lambda x, y: x & y, exit_long_conditions), "exit_long"] = 1
+
+        exit_short_conditions = [df["do_predict"] == 1, df["&-action"] == 4]
+        if exit_short_conditions:
+            df.loc[reduce(lambda x, y: x & y, exit_short_conditions), "exit_short"] = 1
+
+        return df
+```
+
+It is important to consider that `&-action` depends on which environment they choose to use. The example above shows 5 actions, where 0 is neutral, 1 is enter long, 2 is exit long, 3 is enter short and 4 is exit short.
+
+## Configuring the Reinforcement Learner
+
+In order to configure the `Reinforcement Learner` the following dictionary must exist in the `freqai` config:
+
+```json
+        "rl_config": {
+            "train_cycles": 25,
+            "add_state_info": true,
+            "max_trade_duration_candles": 300,
+            "max_training_drawdown_pct": 0.02,
+            "cpu_count": 8,
+            "model_type": "PPO",
+            "policy_type": "MlpPolicy",
+            "model_reward_parameters": {
+                "rr": 1,
+                "profit_aim": 0.025
+            }
+        }
+```
+
+Parameter details can be found [here](freqai-parameter-table.md), but in general the `train_cycles` decides how many times the agent should cycle through the candle data in its artificial environment to train weights in the model. `model_type` is a string which selects one of the available models in [stable_baselines](https://stable-baselines3.readthedocs.io/en/master/)(external link).
+
+!!! Note
+    If you would like to experiment with `continual_learning`, then you should set that value to `true` in the main `freqai` configuration dictionary. This will tell the Reinforcement Learning library to continue training new models from the final state of previous models, instead of retraining new models from scratch each time a retrain is initiated.
+
+!!! Note
+    Remember that the general `model_training_parameters` dictionary should contain all the model hyperparameter customizations for the particular `model_type`. For example, `PPO` parameters can be found [here](https://stable-baselines3.readthedocs.io/en/master/modules/ppo.html).
+
+## Creating a custom reward function
+
+!!! danger "Not for production"
+    Warning!
+    The reward function provided with the Freqtrade source code is a showcase of functionality designed to show/test as many possible environment control features as possible. It is also designed to run quickly on small computers. This is a benchmark, it is *not* for live production. Please beware that you will need to create your own custom_reward() function or use a template built by other users outside of the Freqtrade source code.
+
+As you begin to modify the strategy and the prediction model, you will quickly realize some important differences between the Reinforcement Learner and the Regressors/Classifiers. Firstly, the strategy does not set a target value (no labels!). Instead, you set the `calculate_reward()` function inside the `MyRLEnv` class (see below). A default `calculate_reward()` is provided inside `prediction_models/ReinforcementLearner.py` to demonstrate the necessary building blocks for creating rewards, but this is *not* designed for production. Users *must* create their own custom reinforcement learning model class or use a pre-built one from outside the Freqtrade source code and save it to `user_data/freqaimodels`. It is inside the `calculate_reward()` where creative theories about the market can be expressed. For example, you can reward your agent when it makes a winning trade, and penalize the agent when it makes a losing trade. Or perhaps, you wish to reward the agent for entering trades, and penalize the agent for sitting in trades too long. Below we show examples of how these rewards are all calculated:
+
+!!! note "Hint"
+    The best reward functions are ones that are continuously differentiable, and well scaled. In other words, adding a single large negative penalty to a rare event is not a good idea, and the neural net will not be able to learn that function. Instead, it is better to add a small negative penalty to a common event. This will help the agent learn faster. Not only this, but you can help improve the continuity of your rewards/penalties by having them scale with severity according to some linear/exponential functions. In other words, you'd slowly scale the penalty as the duration of the trade increases. This is better than a single large penalty occurring at a single point in time.
+
+```python
+from freqtrade.freqai.prediction_models.ReinforcementLearner import ReinforcementLearner
+from freqtrade.freqai.RL.Base5ActionRLEnv import Actions, Base5ActionRLEnv, Positions
+
+
+class MyCoolRLModel(ReinforcementLearner):
+    """
+    User created RL prediction model.
+
+    Save this file to `freqtrade/user_data/freqaimodels`
+
+    then use it with:
+
+    freqtrade trade --freqaimodel MyCoolRLModel --config config.json --strategy SomeCoolStrat
+
+    Here the users can override any of the functions
+    available in the `IFreqaiModel` inheritance tree. Most importantly for RL, this
+    is where the user overrides `MyRLEnv` (see below), to define custom
+    `calculate_reward()` function, or to override any other parts of the environment.
+
+    This class also allows users to override any other part of the IFreqaiModel tree.
+    For example, the user can override `def fit()` or `def train()` or `def predict()`
+    to take fine-tuned control over these processes.
+
+    Another common override may be `def data_cleaning_predict()` where the user can
+    take fine-tuned control over the data handling pipeline.
+    """
+    class MyRLEnv(Base5ActionRLEnv):
+        """
+        User made custom environment. This class inherits from BaseEnvironment and gym.Env.
+        Users can override any functions from those parent classes. Here is an example
+        of a user customized `calculate_reward()` function.
+
+        Warning!
+        This is function is a showcase of functionality designed to show as many possible
+        environment control features as possible. It is also designed to run quickly
+        on small computers. This is a benchmark, it is *not* for live production.
+        """
+        def calculate_reward(self, action: int) -> float:
+            # first, penalize if the action is not valid
+            if not self._is_valid(action):
+                return -2
+            pnl = self.get_unrealized_profit()
+
+            factor = 100
+
+            pair = self.pair.replace(':', '')
+
+            # you can use feature values from dataframe
+            # Assumes the shifted RSI indicator has been generated in the strategy.
+            rsi_now = self.raw_features[f"%-rsi-period_10_shift-1_{pair}_"
+                            f"{self.config['timeframe']}"].iloc[self._current_tick]
+
+            # reward agent for entering trades
+            if (action in (Actions.Long_enter.value, Actions.Short_enter.value)
+                    and self._position == Positions.Neutral):
+                if rsi_now < 40:
+                    factor = 40 / rsi_now
+                else:
+                    factor = 1
+                return 25 * factor
+
+            # discourage agent from not entering trades
+            if action == Actions.Neutral.value and self._position == Positions.Neutral:
+                return -1
+            max_trade_duration = self.rl_config.get('max_trade_duration_candles', 300)
+            trade_duration = self._current_tick - self._last_trade_tick
+            if trade_duration <= max_trade_duration:
+                factor *= 1.5
+            elif trade_duration > max_trade_duration:
+                factor *= 0.5
+            # discourage sitting in position
+            if self._position in (Positions.Short, Positions.Long) and \
+            action == Actions.Neutral.value:
+                return -1 * trade_duration / max_trade_duration
+            # close long
+            if action == Actions.Long_exit.value and self._position == Positions.Long:
+                if pnl > self.profit_aim * self.rr:
+                    factor *= self.rl_config['model_reward_parameters'].get('win_reward_factor', 2)
+                return float(pnl * factor)
+            # close short
+            if action == Actions.Short_exit.value and self._position == Positions.Short:
+                if pnl > self.profit_aim * self.rr:
+                    factor *= self.rl_config['model_reward_parameters'].get('win_reward_factor', 2)
+                return float(pnl * factor)
+            return 0.
+```
+
+## Using Tensorboard
+
+Reinforcement Learning models benefit from tracking training metrics. FreqAI has integrated Tensorboard to allow users to track training and evaluation performance across all coins and across all retrainings. Tensorboard is activated via the following command:
+
+```bash
+tensorboard --logdir user_data/models/unique-id
+```
+
+where `unique-id` is the `identifier` set in the `freqai` configuration file. This command must be run in a separate shell to view the output in the browser at 127.0.0.1:6006 (6006 is the default port used by Tensorboard).
+
+![tensorboard](assets/tensorboard.jpg)
+
+## Custom logging
+
+FreqAI also provides a built in episodic summary logger called `self.tensorboard_log` for adding custom information to the Tensorboard log. By default, this function is already called once per step inside the environment to record the agent actions. All values accumulated for all steps in a single episode are reported at the conclusion of each episode, followed by a full reset of all metrics to 0 in preparation for the subsequent episode.
+
+`self.tensorboard_log` can also be used anywhere inside the environment, for example, it can be added to the `calculate_reward` function to collect more detailed information about how often various parts of the reward were called:
+
+```python
+    class MyRLEnv(Base5ActionRLEnv):
+        """
+        User made custom environment. This class inherits from BaseEnvironment and gym.Env.
+        Users can override any functions from those parent classes. Here is an example
+        of a user customized `calculate_reward()` function.
+        """
+        def calculate_reward(self, action: int) -> float:
+            if not self._is_valid(action):
+                self.tensorboard_log("invalid")
+                return -2
+
+```
+
+!!! Note
+    The `self.tensorboard_log()` function is designed for tracking incremented objects only i.e. events, actions inside the training environment. If the event of interest is a float, the float can be passed as the second argument e.g. `self.tensorboard_log("float_metric1", 0.23)`. In this case the metric values are not incremented.
+
+## Choosing a base environment
+
+FreqAI provides three base environments, `Base3ActionRLEnvironment`, `Base4ActionEnvironment` and `Base5ActionEnvironment`. As the names imply, the environments are customized for agents that can select from 3, 4 or 5 actions. The `Base3ActionEnvironment` is the simplest, the agent can select from hold, long, or short. This environment can also be used for long-only bots (it automatically follows the `can_short` flag from the strategy), where long is the enter condition and short is the exit condition. Meanwhile, in the `Base4ActionEnvironment`, the agent can enter long, enter short, hold neutral, or exit position. Finally, in the `Base5ActionEnvironment`, the agent has the same actions as Base4, but instead of a single exit action, it separates exit long and exit short. The main changes stemming from the environment selection include:
+
+* the actions available in the `calculate_reward`
+* the actions consumed by the user strategy
+
+All of the FreqAI provided environments inherit from an action/position agnostic environment object called the `BaseEnvironment`, which contains all shared logic. The architecture is designed to be easily customized. The simplest customization is the `calculate_reward()` (see details [here](#creating-a-custom-reward-function)). However, the customizations can be further extended into any of the functions inside the environment. You can do this by simply overriding those functions inside your `MyRLEnv` in the prediction model file. Or for more advanced customizations, it is encouraged to create an entirely new environment inherited from `BaseEnvironment`.
+
+!!! Note
+    Only the `Base3ActionRLEnv` can do long-only training/trading (set the user strategy attribute `can_short = False`).

md/freqai-running.md

@@ -0,0 +1,187 @@
+# Running FreqAI
+
+There are two ways to train and deploy an adaptive machine learning model - live deployment and historical backtesting. In both cases, FreqAI runs/simulates periodic retraining of models as shown in the following figure:
+
+![freqai-window](assets/freqai_moving-window.jpg)
+
+## Live deployments
+
+FreqAI can be run dry/live using the following command:
+
+```bash
+freqtrade trade --strategy FreqaiExampleStrategy --config config_freqai.example.json --freqaimodel LightGBMRegressor
+```
+
+When launched, FreqAI will start training a new model, with a new `identifier`, based on the config settings. Following training, the model will be used to make predictions on incoming candles until a new model is available. New models are typically generated as often as possible, with FreqAI managing an internal queue of the coin pairs to try to keep all models equally up to date. FreqAI will always use the most recently trained model to make predictions on incoming live data. If you do not want FreqAI to retrain new models as often as possible, you can set `live_retrain_hours` to tell FreqAI to wait at least that number of hours before training a new model. Additionally, you can set `expired_hours` to tell FreqAI to avoid making predictions on models that are older than that number of hours.
+
+Trained models are by default saved to disk to allow for reuse during backtesting or after a crash. You can opt to [purge old models](#purging-old-model-data) to save disk space by setting `"purge_old_models": true` in the config.
+
+To start a dry/live run from a saved backtest model (or from a previously crashed dry/live session), you only need to specify the `identifier` of the specific model:
+
+```json
+    "freqai": {
+        "identifier": "example",
+        "live_retrain_hours": 0.5
+    }
+```
+
+In this case, although FreqAI will initiate with a pre-trained model, it will still check to see how much time has elapsed since the model was trained. If a full `live_retrain_hours` has elapsed since the end of the loaded model, FreqAI will start training a new model.
+
+### Automatic data download
+
+FreqAI automatically downloads the proper amount of data needed to ensure training of a model through the defined `train_period_days` and `startup_candle_count` (see the [parameter table](freqai-parameter-table.md) for detailed descriptions of these parameters). 
+
+### Saving prediction data
+
+All predictions made during the lifetime of a specific `identifier` model are stored in `historic_predictions.pkl` to allow for reloading after a crash or changes made to the config.
+
+### Purging old model data
+
+FreqAI stores new model files after each successful training. These files become obsolete as new models are generated to adapt to new market conditions. If you are planning to leave FreqAI running for extended periods of time with high frequency retraining, you should enable `purge_old_models` in the config:
+
+```json
+    "freqai": {
+        "purge_old_models": 4,
+    }
+```
+
+This will automatically purge all models older than the four most recently trained ones to save disk space. Inputing "0" will never purge any models.
+
+## Backtesting
+
+The FreqAI backtesting module can be executed with the following command:
+
+```bash
+freqtrade backtesting --strategy FreqaiExampleStrategy --strategy-path freqtrade/templates --config config_examples/config_freqai.example.json --freqaimodel LightGBMRegressor --timerange 20210501-20210701
+```
+
+If this command has never been executed with the existing config file, FreqAI will train a new model
+for each pair, for each backtesting window within the expanded `--timerange`.
+
+Backtesting mode requires [downloading the necessary data](#downloading-data-to-cover-the-full-backtest-period) before deployment (unlike in dry/live mode where FreqAI handles the data downloading automatically). You should be careful to consider that the time range of the downloaded data is more than the backtesting time range. This is because FreqAI needs data prior to the desired backtesting time range in order to train a model to be ready to make predictions on the first candle of the set backtesting time range. More details on how to calculate the data to download can be found [here](#deciding-the-size-of-the-sliding-training-window-and-backtesting-duration).
+
+!!! Note "Model reuse"
+    Once the training is completed, you can execute the backtesting again with the same config file and
+    FreqAI will find the trained models and load them instead of spending time training. This is useful
+    if you want to tweak (or even hyperopt) buy and sell criteria inside the strategy. If you
+    *want* to retrain a new model with the same config file, you should simply change the `identifier`.
+    This way, you can return to using any model you wish by simply specifying the `identifier`.
+
+!!! Note
+    Backtesting calls `set_freqai_targets()` one time for each backtest window (where the number of windows is the full backtest timerange divided by the `backtest_period_days` parameter). Doing this means that the targets simulate dry/live behavior without look ahead bias. However, the definition of the features in `feature_engineering_*()` is performed once on the entire training timerange. This means that you should be sure that features do not look-ahead into the future.
+    More details about look-ahead bias can be found in [Common Mistakes](strategy-customization.md#common-mistakes-when-developing-strategies).
+
+---
+
+### Saving backtesting prediction data
+
+To allow for tweaking your strategy (**not** the features!), FreqAI will automatically save the predictions during backtesting so that they can be reused for future backtests and live runs using the same `identifier` model. This provides a performance enhancement geared towards enabling **high-level hyperopting** of entry/exit criteria.
+
+An additional directory called `backtesting_predictions`, which contains all the predictions stored in `feather` format, will be created in the `unique-id` folder.
+
+To change your **features**, you **must** set a new `identifier` in the config to signal to FreqAI to train new models.
+
+To save the models generated during a particular backtest so that you can start a live deployment from one of them instead of training a new model, you must set `save_backtest_models` to `True` in the config.
+
+!!! Note
+    To ensure that the model can be reused, freqAI will call your strategy with a dataframe of length 1. 
+    If your strategy requires more data than this to generate the same features, you can't reuse backtest predictions for live deployment and need to update your `identifier` for each new backtest.
+
+### Backtest live collected predictions
+
+FreqAI allow you to reuse live historic predictions through the backtest parameter `--freqai-backtest-live-models`. This can be useful when you want to reuse predictions generated in dry/run for comparison or other study.
+
+The `--timerange` parameter must not be informed, as it will be automatically calculated through the data in the historic predictions file.
+
+### Downloading data to cover the full backtest period
+
+For live/dry deployments, FreqAI will download the necessary data automatically. However, to use backtesting functionality, you need to download the necessary data using `download-data` (details [here](data-download.md#data-downloading)). You need to pay careful attention to understanding how much *additional* data needs to be downloaded to ensure that there is a sufficient amount of training data *before* the start of the backtesting time range. The amount of additional data can be roughly estimated by moving the start date of the time range backwards by `train_period_days` and the `startup_candle_count` (see the [parameter table](freqai-parameter-table.md) for detailed descriptions of these parameters) from the beginning of the desired backtesting time range. 
+
+As an example, to backtest the `--timerange 20210501-20210701` using the [example config](freqai-configuration.md#setting-up-the-configuration-file) which sets `train_period_days` to 30, together with `startup_candle_count: 40` on a maximum `include_timeframes` of 1h, the start date for the downloaded data needs to be `20210501` - 30 days - 40 * 1h / 24 hours = 20210330 (31.7 days earlier than the start of the desired training time range).
+
+### Deciding the size of the sliding training window and backtesting duration
+
+The backtesting time range is defined with the typical `--timerange` parameter in the configuration file. The duration of the sliding training window is set by `train_period_days`, whilst `backtest_period_days` is the sliding backtesting window, both in number of days (`backtest_period_days` can be
+a float to indicate sub-daily retraining in live/dry mode). In the presented [example config](freqai-configuration.md#setting-up-the-configuration-file) (found in `config_examples/config_freqai.example.json`), the user is asking FreqAI to use a training period of 30 days and backtest on the subsequent 7 days. After the training of the model, FreqAI will backtest the subsequent 7 days. The "sliding window" then moves one week forward (emulating FreqAI retraining once per week in live mode) and the new model uses the previous 30 days (including the 7 days used for backtesting by the previous model) to train. This is repeated until the end of `--timerange`.  This means that if you set `--timerange 20210501-20210701`, FreqAI will have trained 8 separate models at the end of `--timerange` (because the full range comprises 8 weeks).
+
+!!! Note
+    Although fractional `backtest_period_days` is allowed, you should be aware that the `--timerange` is divided by this value to determine the number of models that FreqAI will need to train in order to backtest the full range. For example, by setting a `--timerange` of 10 days, and a `backtest_period_days` of 0.1, FreqAI will need to train 100 models per pair to complete the full backtest. Because of this, a true backtest of FreqAI adaptive training would take a *very* long time. The best way to fully test a model is to run it dry and let it train constantly. In this case, backtesting would take the exact same amount of time as a dry run.
+
+## Defining model expirations
+
+During dry/live mode, FreqAI trains each coin pair sequentially (on separate threads/GPU from the main Freqtrade bot). This means that there is always an age discrepancy between models. If you are training on 50 pairs, and each pair requires 5 minutes to train, the oldest model will be over 4 hours old. This may be undesirable if the characteristic time scale (the trade duration target) for a strategy is less than 4 hours. You can decide to only make trade entries if the model is less than a certain number of hours old by setting the `expiration_hours` in the config file:
+
+```json
+    "freqai": {
+        "expiration_hours": 0.5,
+    }
+```
+
+In the presented example config, the user will only allow predictions on models that are less than 1/2 hours old.
+
+## Controlling the model learning process
+
+Model training parameters are unique to the selected machine learning library. FreqAI allows you to set any parameter for any library using the `model_training_parameters` dictionary in the config. The example config (found in `config_examples/config_freqai.example.json`) shows some of the example parameters associated with `Catboost` and `LightGBM`, but you can add any parameters available in those libraries or any other machine learning library you choose to implement.
+
+Data split parameters are defined in `data_split_parameters` which can be any parameters associated with scikit-learn's `train_test_split()` function. `train_test_split()` has a parameters called `shuffle` which allows to shuffle the data or keep it unshuffled. This is particularly useful to avoid biasing training with temporally auto-correlated data. More details about these parameters can be found the [scikit-learn website](https://scikit-learn.org/stable/modules/generated/sklearn.model_selection.train_test_split.html) (external website).
+
+The FreqAI specific parameter `label_period_candles` defines the offset (number of candles into the future) used for the `labels`. In the presented [example config](freqai-configuration.md#setting-up-the-configuration-file), the user is asking for `labels` that are 24 candles in the future.
+
+## Continual learning
+
+You can choose to adopt a continual learning scheme by setting `"continual_learning": true` in the config. By enabling `continual_learning`, after training an initial model from scratch, subsequent trainings will start from the final model state of the preceding training. This gives the new model a "memory" of the previous state. By default, this is set to `False` which means that all new models are trained from scratch, without input from previous models.
+
+???+ danger "Continual learning enforces a constant parameter space"
+    Since `continual_learning` means that the model parameter space *cannot* change between trainings, `principal_component_analysis` is automatically disabled when `continual_learning` is enabled. Hint: PCA changes the parameter space and the number of features, learn more about PCA [here](freqai-feature-engineering.md#data-dimensionality-reduction-with-principal-component-analysis).
+
+???+ danger "Experimental functionality"
+    Beware that this is currently a naive approach to incremental learning, and it has a high probability of overfitting/getting stuck in local minima while the market moves away from your model. We have the mechanics available in FreqAI primarily for experimental purposes and so that it is ready for more mature approaches to continual learning in chaotic systems like the crypto market.
+
+## Hyperopt
+
+You can hyperopt using the same command as for [typical Freqtrade hyperopt](hyperopt.md):
+
+```bash
+freqtrade hyperopt --hyperopt-loss SharpeHyperOptLoss --strategy FreqaiExampleStrategy --freqaimodel LightGBMRegressor --strategy-path freqtrade/templates --config config_examples/config_freqai.example.json --timerange 20220428-20220507
+```
+
+`hyperopt` requires you to have the data pre-downloaded in the same fashion as if you were doing [backtesting](#backtesting). In addition, you must consider some restrictions when trying to hyperopt FreqAI strategies:
+
+- The `--analyze-per-epoch` hyperopt parameter is not compatible with FreqAI.
+- It's not possible to hyperopt indicators in the `feature_engineering_*()` and `set_freqai_targets()` functions. This means that you cannot optimize model parameters using hyperopt. Apart from this exception, it is possible to optimize all other [spaces](hyperopt.md#running-hyperopt-with-smaller-search-space).
+- The backtesting instructions also apply to hyperopt.
+
+The best method for combining hyperopt and FreqAI is to focus on hyperopting entry/exit thresholds/criteria. You need to focus on hyperopting parameters that are not used in your features. For example, you should not try to hyperopt rolling window lengths in the feature creation, or any part of the FreqAI config which changes predictions. In order to efficiently hyperopt the FreqAI strategy, FreqAI stores predictions as dataframes and reuses them. Hence the requirement to hyperopt entry/exit thresholds/criteria only.
+
+A good example of a hyperoptable parameter in FreqAI is a threshold for the [Dissimilarity Index (DI)](freqai-feature-engineering.md#identifying-outliers-with-the-dissimilarity-index-di) `DI_values` beyond which we consider data points as outliers:
+
+```python
+di_max = IntParameter(low=1, high=20, default=10, space='buy', optimize=True, load=True)
+dataframe['outlier'] = np.where(dataframe['DI_values'] > self.di_max.value/10, 1, 0)
+```
+
+This specific hyperopt would help you understand the appropriate `DI_values` for your particular parameter space.
+
+## Using Tensorboard
+
+!!! note "Availability"
+    FreqAI includes tensorboard for a variety of models, including XGBoost, all PyTorch models, Reinforcement Learning, and Catboost. If you would like to see Tensorboard integrated into another model type, please open an issue on the [Freqtrade GitHub](https://github.com/freqtrade/freqtrade/issues)
+
+!!! danger "Requirements"
+    Tensorboard logging requires the FreqAI torch installation/docker image.
+
+
+The easiest way to use tensorboard is to ensure `freqai.activate_tensorboard` is set to `True` (default setting) in your configuration file, run FreqAI, then open a separate shell and run:
+
+```bash
+cd freqtrade
+tensorboard --logdir user_data/models/unique-id
+```
+
+where `unique-id` is the `identifier` set in the `freqai` configuration file. This command must be run in a separate shell if you wish to view the output in your browser at 127.0.0.1:6060 (6060 is the default port used by Tensorboard).
+
+![tensorboard](assets/tensorboard.jpg)
+
+
+!!! note "Deactivate for improved performance"
+    Tensorboard logging can slow down training and should be deactivated for production use.

md/freqai.md

@@ -0,0 +1,141 @@
+![freqai-logo](assets/freqai_doc_logo.svg)
+
+# FreqAI
+
+## Introduction
+
+FreqAI is a software designed to automate a variety of tasks associated with training a predictive machine learning model to generate market forecasts given a set of input signals. In general, FreqAI aims to be a sandbox for easily deploying robust machine learning libraries on real-time data ([details](#freqai-position-in-open-source-machine-learning-landscape)).
+
+!!! Note
+    FreqAI is, and always will be, a not-for-profit, open-source project. FreqAI does *not* have a crypto token, FreqAI does *not* sell signals, and FreqAI does not have a domain besides the present [freqtrade documentation](https://www.freqtrade.io/en/latest/freqai/).
+
+Features include:
+
+* **Self-adaptive retraining** - Retrain models during [live deployments](freqai-running.md#live-deployments) to self-adapt to the market in a supervised manner
+* **Rapid feature engineering** - Create large rich [feature sets](freqai-feature-engineering.md#feature-engineering) (10k+ features) based on simple user-created strategies
+* **High performance** - Threading allows for adaptive model retraining on a separate thread (or on GPU if available) from model inferencing (prediction) and bot trade operations. Newest models and data are kept in RAM for rapid inferencing
+* **Realistic backtesting** - Emulate self-adaptive training on historic data with a [backtesting module](freqai-running.md#backtesting) that automates retraining
+* **Extensibility** - The generalized and robust architecture allows for incorporating any [machine learning library/method](freqai-configuration.md#using-different-prediction-models) available in Python. Eight examples are currently available, including classifiers, regressors, and a convolutional neural network
+* **Smart outlier removal** - Remove outliers from training and prediction data sets using a variety of [outlier detection techniques](freqai-feature-engineering.md#outlier-detection)
+* **Crash resilience** - Store trained models to disk to make reloading from a crash fast and easy, and [purge obsolete files](freqai-running.md#purging-old-model-data) for sustained dry/live runs
+* **Automatic data normalization** - [Normalize the data](freqai-feature-engineering.md#building-the-data-pipeline) in a smart and statistically safe way
+* **Automatic data download** - Compute timeranges for data downloads and update historic data (in live deployments)
+* **Cleaning of incoming data** - Handle NaNs safely before training and model inferencing
+* **Dimensionality reduction** - Reduce the size of the training data via [Principal Component Analysis](freqai-feature-engineering.md#data-dimensionality-reduction-with-principal-component-analysis)
+* **Deploying bot fleets** - Set one bot to train models while a fleet of [consumers](producer-consumer.md) use signals.
+
+## Quick start
+
+The easiest way to quickly test FreqAI is to run it in dry mode with the following command:
+
+```bash
+freqtrade trade --config config_examples/config_freqai.example.json --strategy FreqaiExampleStrategy --freqaimodel LightGBMRegressor --strategy-path freqtrade/templates
+```
+
+You will see the boot-up process of automatic data downloading, followed by simultaneous training and trading. 
+
+!!! danger "Not for production"
+    The example strategy provided with the Freqtrade source code is designed for showcasing/testing a wide variety of FreqAI features. It is also designed to run on small computers so that it can be used as a benchmark between developers and users. It is *not* designed to be run in production.
+
+An example strategy, prediction model, and config to use as a starting points can be found in
+`freqtrade/templates/FreqaiExampleStrategy.py`, `freqtrade/freqai/prediction_models/LightGBMRegressor.py`, and
+`config_examples/config_freqai.example.json`, respectively.
+
+## General approach
+
+You provide FreqAI with a set of custom *base indicators* (the same way as in a [typical Freqtrade strategy](strategy-customization.md)) as well as target values (*labels*). For each pair in the whitelist, FreqAI trains a model to predict the target values based on the input of custom indicators. The models are then consistently retrained, with a predetermined frequency, to adapt to market conditions. FreqAI offers the ability to both backtest strategies (emulating reality with periodic retraining on historic data) and deploy dry/live runs. In dry/live conditions, FreqAI can be set to constant retraining in a background thread to keep models as up to date as possible.
+
+An overview of the algorithm, explaining the data processing pipeline and model usage, is shown below.
+
+![freqai-algo](assets/freqai_algo.jpg)
+
+### Important machine learning vocabulary
+
+**Features** - the parameters, based on historic data, on which a model is trained. All features for a single candle are stored as a vector. In FreqAI, you build a feature data set from anything you can construct in the strategy.
+
+**Labels** - the target values that the model is trained toward. Each feature vector is associated with a single label that is defined by you within the strategy. These labels intentionally look into the future and are what you are training the model to be able to predict.
+
+**Training** - the process of "teaching" the model to match the feature sets to the associated labels. Different types of models "learn" in different ways which means that one might be better than another for a specific application. More information about the different models that are already implemented in FreqAI can be found [here](freqai-configuration.md#using-different-prediction-models).
+
+**Train data** - a subset of the feature data set that is fed to the model during training to "teach" the model how to predict the targets. This data directly influences weight connections in the model.
+
+**Test data** - a subset of the feature data set that is used to evaluate the performance of the model after training. This data does not influence nodal weights within the model.
+
+**Inferencing** - the process of feeding a trained model new unseen data on which it will make a prediction. 
+
+## Install prerequisites
+
+The normal Freqtrade install process will ask if you wish to install FreqAI dependencies. You should reply "yes" to this question if you wish to use FreqAI. If you did not reply yes, you can manually install these dependencies after the install with:
+
+``` bash
+pip install -r requirements-freqai.txt
+```
+
+!!! Note
+    Catboost will not be installed on low-powered arm devices (raspberry), since it does not provide wheels for this platform.
+
+### Usage with docker
+
+If you are using docker, a dedicated tag with FreqAI dependencies is available as `:freqai`. As such - you can replace the image line in your docker compose file with `image: freqtradeorg/freqtrade:stable_freqai`. This image contains the regular FreqAI dependencies. Similar to native installs, Catboost will not be available on ARM based devices. If you would like to use PyTorch or Reinforcement learning, you should use the torch or RL tags, `image: freqtradeorg/freqtrade:stable_freqaitorch`, `image: freqtradeorg/freqtrade:stable_freqairl`.
+
+!!! note "docker-compose-freqai.yml"
+   We do provide an explicit docker-compose file for this in `docker/docker-compose-freqai.yml` - which can be used via `docker compose -f docker/docker-compose-freqai.yml run ...` - or can be copied to replace the original docker file. This docker-compose file also contains a (disabled) section to enable GPU resources within docker containers. This obviously assumes the system has GPU resources available.
+
+### FreqAI position in open-source machine learning landscape
+
+Forecasting chaotic time-series based systems, such as equity/cryptocurrency markets, requires a broad set of tools geared toward testing a wide range of hypotheses. Fortunately, a recent maturation of robust machine learning libraries (e.g. `scikit-learn`) has opened up a wide range of research possibilities. Scientists from a diverse range of fields can now easily prototype their studies on an abundance of established machine learning algorithms. Similarly, these user-friendly libraries enable "citizen scientists" to use their basic Python skills for data exploration. However, leveraging these machine learning libraries on historical and live chaotic data sources can be logistically difficult and expensive. Additionally, robust data collection, storage, and handling presents a disparate challenge. [`FreqAI`](#freqai) aims to provide a generalized and extensible open-sourced framework geared toward live deployments of adaptive modeling for market forecasting. The `FreqAI` framework is effectively a sandbox for the rich world of open-source machine learning libraries. Inside the `FreqAI` sandbox, users find they can combine a wide variety of third-party libraries to test creative hypotheses on a free live 24/7 chaotic data source - cryptocurrency exchange data. 
+
+### Citing FreqAI
+
+FreqAI is [published in the Journal of Open Source Software](https://joss.theoj.org/papers/10.21105/joss.04864). If you find FreqAI useful in your research, please use the following citation:
+
+```bibtex
+@article{Caulk2022, 
+    doi = {10.21105/joss.04864},
+    url = {https://doi.org/10.21105/joss.04864},
+    year = {2022}, publisher = {The Open Journal},
+    volume = {7}, number = {80}, pages = {4864},
+    author = {Robert A. Caulk and Elin Törnquist and Matthias Voppichler and Andrew R. Lawless and Ryan McMullan and Wagner Costa Santos and Timothy C. Pogue and Johan van der Vlugt and Stefan P. Gehring and Pascal Schmidt},
+    title = {FreqAI: generalizing adaptive modeling for chaotic time-series market forecasts},
+    journal = {Journal of Open Source Software} } 
+```
+
+## Common pitfalls
+
+FreqAI cannot be combined with dynamic `VolumePairlists` (or any pairlist filter that adds and removes pairs dynamically).
+This is for performance reasons - FreqAI relies on making quick predictions/retrains. To do this effectively,
+it needs to download all the training data at the beginning of a dry/live instance. FreqAI stores and appends
+new candles automatically for future retrains. This means that if new pairs arrive later in the dry run due to a volume pairlist, it will not have the data ready. However, FreqAI does work with the `ShufflePairlist` or a `VolumePairlist` which keeps the total pairlist constant (but reorders the pairs according to volume).
+
+## Additional learning materials
+
+Here we compile some external materials that provide deeper looks into various components of FreqAI:
+
+- [Real-time head-to-head: Adaptive modeling of financial market data using XGBoost and CatBoost](https://emergentmethods.medium.com/real-time-head-to-head-adaptive-modeling-of-financial-market-data-using-xgboost-and-catboost-995a115a7495)
+- [FreqAI - from price to prediction](https://emergentmethods.medium.com/freqai-from-price-to-prediction-6fadac18b665)
+
+
+## Support
+
+You can find support for FreqAI in a variety of places, including the [Freqtrade discord](https://discord.gg/Jd8JYeWHc4), the dedicated [FreqAI discord](https://discord.gg/7AMWACmbjT), and in [github issues](https://github.com/freqtrade/freqtrade/issues).
+
+## Credits
+
+FreqAI is developed by a group of individuals who all contribute specific skillsets to the project.
+
+Conception and software development:
+Robert Caulk @robcaulk
+
+Theoretical brainstorming and data analysis:
+Elin Törnquist @th0rntwig
+
+Code review and software architecture brainstorming:
+@xmatthias
+
+Software development:
+Wagner Costa @wagnercosta
+Emre Suzen @aemr3
+Timothy Pogue @wizrds
+
+Beta testing and bug reporting:
+Stefan Gehring @bloodhunter4rc, @longyu, Andrew Lawless @paranoidandy, Pascal Schmidt @smidelis, Ryan McMullan @smarmau, Juha Nykänen @suikula, Johan van der Vlugt @jooopiert, Richárd Józsa @richardjosza

md/plan.md

@@ -0,0 +1,536 @@
+### Freqtrade AI 策略设计文档（补充完善版）
+
+设计目标
+
+本策略旨在基于 Freqtrade 平台，设计一个利用 CCXT 绑定 OKX 交易所数据的 AI 交易策略，仅进行多头交易（不做空），使用回归模型进行预测，充分利用 AI 模型优化，确保运行可扩展、可追踪、可评估，代码可维护性强，并通过循序渐进的优化实现稳定持续的盈利预期。具体目标包括：
+
+1. 充分利用 AI 模型进行优化：基于回归模型，优化多头交易信号生成。
+    
+2. AI 模型运行可扩展、可追踪、可评估：支持模型切换、版本管理，记录详细性能指标。
+    
+3. 代码可维护性强：通过模块化设计和抽象接口提高代码可读性和扩展性。
+    
+4. 循序渐进优化：从简单模型开始，逐步引入复杂特征，最终实现稳定盈利。
+    
+5. 数据来源专一性（新增）：所有数据通过 CCXT 从 OKX 交易所获取，确保数据一致性和可靠性。
+    
+6. 仅做多头交易（新增）：策略仅生成买入和持有信号，不涉及卖空操作。
+    
+7. 回归模型（新增）：使用回归模型预测价格变化或收益率，生成交易信号。
+    
+
+---
+
+标签列的明确与初始化
+
+标签列分类
+
+所有标签列基于 CCXT 从 OKX 交易所获取的数据（如 K 线、成交量、订单簿、资金费率等），分为以下几类：
+
+目标变量（Target Variables）
+
+- &-s_close：未来一段时间内价格变化（收盘价差值），用于回归模型训练和预测。
+    
+- &-roi_0, &-roi_15, &-roi_30：不同时间窗口下的收益率目标值（如 0、15、30 分钟），用于动态设置止盈。
+    
+- 调整（新增）：移除 &-up_or_down（二分类目标），因回归模型仅需连续值目标。
+    
+
+特征列（Feature Columns）
+
+- %-%-rsi-period, %-%-macd-period, %-%-bb_width-period：基于 OKX K 线数据的扩展技术指标（如 RSI、MACD、布林带宽度）。
+    
+- %-%-volume_ma：基于 OKX 成交量的移动平均值，判断市场活跃度。
+    
+- %-%-order_book_imbalance（新增）：基于 OKX 订单簿数据的买卖盘不平衡特征，反映市场供需。
+    
+- %-%-funding_rate（新增）：基于 OKX 资金费率数据，反映市场情绪，辅助多头信号生成。
+    
+- 补充：若 OKX API 支持，可引入链上数据（如链上转账量、活跃地址数），提升特征多样性。
+    
+
+预测结果列（Prediction Columns）
+
+- do_predict：预测可信度，取值 1（可信）、0（不可信）、-1（异常数据）。
+    
+- DI_values：Dissimilarity Index（DI）值，识别预测异常点。
+    
+- &-action：回归模型预测的交易动作（买入或持有），不包含卖空动作。
+    
+
+元数据列（Metadata Columns）
+
+- &-buy_rsi_pred：动态计算的 RSI 买入阈值，用于生成多头交易条件。
+    
+- &-stoploss_pred, &-roi_0_pred：动态计算的止损和止盈阈值，优化多头交易策略。
+    
+
+优雅初始化标签列的方法
+
+为确保鲁棒性，所有标签列在 set_freqai_targets 方法中初始化，基于 OKX 数据并遵循以下原则：
+
+统一命名规范
+
+- 目标变量以 &- 开头，确保 FreqAI 识别。
+    
+- 特征列以 %-%- 开头，支持自动扩展。
+    
+- 预测结果列以 &- 或 %- 开头，确保正确处理。
+    
+
+初始化逻辑清晰
+
+- 检查 DataFrame 是否包含 OKX 数据必要列（如 close、volume、order_book、funding_rate），若缺失则抛出异常或填充默认值。
+    
+- 目标变量（如 &-s_close）基于 OKX K 线未来价格变化计算，符合回归模型要求。
+    
+- 特征列（如 %-%-rsi-period）通过 feature_engineering_expand_all 扩展，确保基于 OKX 数据生成。
+    
+- 补充（新增）：验证 OKX 数据时间戳一致性，避免因 API 延迟导致数据错位。
+    
+
+数据清理与填充
+
+- 处理 NaN 和 Inf 值，使用 fillna 或 replace 填充缺失值。
+    
+- 补充（新增）：为 OKX 订单簿或资金费率数据设计备用填充方案（如历史均值），应对数据不可用情况。
+    
+
+日志记录与调试支持
+
+- 输出 DataFrame 列名、形状和前几行 OKX 数据，便于调试。
+    
+- 若某列未生成，抛出异常或记录警告。
+    
+- 补充（新增）：记录 OKX 数据获取的 API 调用状态（如响应时间、错误码），追踪数据源可靠性。
+    
+
+---
+
+增强程序鲁棒性的措施
+
+标签列的健壮性校验
+
+- 校验关键列（如 &-s_close, &-action）基于 OKX 数据生成，若缺失抛出异常。
+    
+- 补充（新增）：校验 OKX 数据完整性（如 K 线连续性、订单簿深度），确保特征和目标变量可靠。
+    
+
+特征工程的可扩展性
+
+- 通过 feature_engineering_expand_all 和 feature_engineering_expand_basic 扩展 OKX 数据特征，支持多时间周期。
+    
+- 在配置文件中定义 include_timeframes 和 include_corr_pairlist，控制特征扩展。
+    
+- 补充（新增）：使用回归模型的特征重要性筛选（如 XGBoost 的 feature importance），减少 OKX 数据衍生特征的冗余，降低过拟合风险。
+    
+
+模型训练与预测的分离
+
+- 训练和预测逻辑分离，确保回归模型训练不干扰多头信号生成。
+    
+- 补充（新增）：缓存 OKX 数据预测结果，优化高频交易场景的计算效率。
+    
+
+模型版本管理与回滚机制
+
+- 使用 identifier 区分回归模型版本，确保保存和加载独立。
+    
+- 若模型表现不佳（如均方误差过大），回滚到上一版本。
+    
+- 补充（新增）：定义回滚触发条件（如验证集 MSE 超过阈值或夏普比率下降），支持多模型并行比较。
+    
+
+计算效率与资源管理
+
+- 优化 OKX 数据处理（如批量加载 K 线、订单簿），减少 CCXT API 调用和计算延迟。
+    
+- 监控 CPU/GPU 使用率，确保实盘信号生成及时。
+    
+- 补充（新增）：记录 OKX API 调用频率，遵守交易所速率限制（如每秒请求数）。
+    
+
+---
+
+防止未来数据泄露的审核机制
+
+特征计算的时序约束
+
+- 所有特征基于 OKX 历史数据（如 K 线、订单簿、资金费率），通过 CCXT 获取，避免未来数据。
+    
+- 在 feature_engineering_expand_all 方法中，校验数据时间戳。
+    
+
+目标变量的时序约束
+
+- 目标变量（如 &-s_close）通过 shift(-label_period_candles) 计算，确保基于未来 OKX 数据但不提前使用。
+    
+- 补充（新增）：审计 OKX 数据时间戳，确保目标变量生成合规。
+    
+
+数据划分的严格控制
+
+- 设置 test_size 和 shuffle，确保训练集和测试集基于 OKX 数据合理划分。
+    
+- 配置 backtest_period_days 和 train_period_days，避免未来数据误用。
+    
+- 补充（新增）：在实盘中监控 OKX API 数据延迟，防止间接泄露。
+    
+
+回测过程的审计日志
+
+- 记录 OKX 数据特征生成、目标变量计算和多头信号生成的时间戳和值。
+    
+- 补充（新增）：生成 OKX 数据审计报告，定期审查回测合规性。
+    
+
+自动化检测机制
+
+- 自动检测 OKX 数据特征和目标变量是否基于历史数据，若发现未来数据，抛出异常。
+    
+- 补充（新增）：实现 OKX 数据时间戳校验工具，扫描潜在泄露点。
+    
+
+---
+
+风险管理与资金管理
+
+仓位与资金分配
+
+- 动态调整仓位，基于 OKX 市场波动性和账户余额，限制单笔交易风险（如账户资金的 2%）。
+    
+- 调整（新增）：仅允许多头仓位，禁止杠杆空头交易，符合只做多要求。
+    
+
+止损与止盈
+
+- 使用动态止损和止盈（如 &-stoploss_pred, &-roi_0_pred），基于回归模型预测调整。
+    
+- 补充（新增）：设计分层止损（如硬性止损 + trailing stop），优化多头盈利，防止过早离场。
+    
+
+极端市场事件处理
+
+- 针对 OKX 市场异常（如价格闪崩、流动性不足），设计保护机制，如暂停交易或降低仓位。
+    
+- 补充（新增）：引入市场压力测试，模拟 OKX 市场极端场景，验证策略稳定性。
+    
+
+---
+
+外部数据源的可靠性和处理
+
+数据源质量控制
+
+- 所有数据通过 CCXT 从 OKX 获取，确保一致性。
+    
+- 对 OKX 订单簿或资金费率数据进行清洗，处理噪声、缺失值或延迟。
+    
+- 补充（新增）：设计备用数据获取方案（如切换 OKX API 端点），应对 API 不可用情况。
+    
+
+数据源合规性
+
+- 确保 OKX 数据使用符合交易所条款和法律法规（如数据隐私）。
+    
+- 补充（新增）：记录 CCXT API 调用和数据使用日志，便于合规性审计。
+    
+
+---
+
+用户交互与监控
+
+实时监控
+
+- 提供实时监控界面，展示回归模型表现、OKX 数据特征和多头信号。
+    
+- 设置自动报警机制，在模型异常或 OKX 市场剧变时通知用户。
+    
+
+交互性设计
+
+- 允许用户调整参数（如止损阈值、仓位比例），提高策略灵活性。
+    
+- 补充（新增）：生成每日/每周策略报告，总结 OKX 市场表现、盈利和优化建议。
+    
+
+---
+
+法律与合规性
+
+- 确保策略符合 OKX 交易所规则和相关法律法规（如加密货币交易合规）。
+    
+- 记录所有交易和 OKX 数据操作，满足审计要求。
+    
+- 补充（新增）：定期与合规专家咨询，确保策略适应 OKX 市场监管变化。
+    
+
+---
+
+策略设计目标的实现路径
+
+充分利用 AI 模型进行优化
+
+- 模型选择：选择回归模型，如 XGBoost、LightGBM 或 PyTorch 回归模型，预测 OKX 价格变化或收益率。
+    
+- 特征工程：基于 OKX 数据，通过 feature_engineering_expand_all 和 feature_engineering_standard 生成丰富特征集。
+    
+- 模型训练：使用 train_test_split 分割 OKX 数据，通过 model_training_parameters 优化超参数。
+    
+- 补充（新增）：定期评估回归模型性能，探索深度学习回归模型（如 LSTM）以捕捉 OKX 市场时间序列特征。
+    
+
+AI 模型运行可扩展、可追踪、可评估
+
+- 可扩展性：通过 freqai.model_training_parameters 支持回归模型切换。
+    
+- 可追踪性：使用 identifier 区分模型版本，确保保存和加载独立。
+    
+- 可评估性：记录均方误差（MSE）、R²、夏普比率、最大回撤等指标，通过 TensorBoard 或日志可视化。
+    
+- 补充（新增）：在 OKX 市场不同条件下（牛市、熊市、震荡市）测试回归模型，确保适应性。
+    
+
+代码可维护性强
+
+- 模块化设计：将 OKX 数据处理、特征工程、回归模型训练、信号生成封装为独立函数或类。
+    
+- 抽象类与接口：继承 BaseRegressorModel，减少重复代码。
+    
+- 配置驱动：通过 config.json 配置模型、特征和训练参数。
+    
+- 补充（新增）：编写 OKX 数据处理和模型训练的文档和测试用例，确保代码易于维护。
+    
+
+循序渐进的优化流程
+
+- 初步验证：使用 OKX 数据和简单回归模型（如 XGBoost）验证多头信号逻辑。
+    
+- 逐步优化：引入复杂特征（如 OKX 订单簿、资金费率）和模型（如 PyTorch 回归）。
+    
+- 持续监控：通过 OKX 数据回测、模拟盘和实盘运行，调整特征和参数。
+    
+- 稳定获利：通过多维度测试，确保策略在 OKX 市场条件下 Profitable。
+    
+- 补充（新增）：设定优化阶段时间表（如初步验证 1 个月，复杂模型引入 3 个月），保持开发节奏。
+    
+
+---
+
+配置建议
+
+以下是符合策略诉求的 Freqtrade 配置建议，基于 OKX 数据、只做多和回归模型：
+
+数据源配置
+
+- exchange:  
+    交易所：
+    
+    json
+    
+    ```json
+    {
+      "exchange": {
+        "name": "okx",
+        "ccxt_config": {
+          "enableRateLimit": true,
+          "rateLimit": 1000  // 遵守 OKX API 速率限制
+        }
+      }
+    }
+    ```
+    
+- pairlists:  
+    pairlists：配对列表
+    
+    - 配置 OKX 交易对（如 BTC/USDT、ETH/USDT），通过 include_corr_pairlist 扩展相关资产。
+        
+    - 示例：
+        
+        json
+        
+        ```json
+        {
+          "pairlists": [
+            {
+              "method": "StaticPairList",
+              "pairs": ["BTC/USDT", "ETH/USDT"]
+            }
+          ]
+        }
+        ```
+        
+
+策略配置
+
+- freqai:  
+    freqai：
+    
+    json
+    
+    ```json
+    {
+      "freqai": {
+        "enabled": true,
+        "identifier": "okx_regression_v1",  // 模型版本标识
+        "model_training_parameters": {
+          "model_type": "XGBoost",  // 初始回归模型，可切换为 LightGBM 或 PyTorch
+          "objective": "reg:squarederror",  // 回归目标
+          "n_estimators": 100,
+          "learning_rate": 0.05
+        },
+        "feature_parameters": {
+          "include_timeframes": ["5m", "15m", "1h"],  // OKX K 线时间周期
+          "include_corr_pairlist": ["BTC/USDT", "ETH/USDT"],
+          "feature_engineering": {
+            "rsi": {"periods": [14, 21]},
+            "macd": {"fast": 12, "slow": 26},
+            "bb_width": {"period": 20},
+            "volume_ma": {"period": 20},
+            "order_book_imbalance": {"depth": 10},  // OKX 订单簿特征
+            "funding_rate": {}  // OKX 资金费率
+          }
+        },
+        "data_split_parameters": {
+          "test_size": 0.2,
+          "shuffle": false  // 防止时间序列数据打乱
+        },
+        "backtest_period_days": 30,
+        "train_period_days": 90
+      }
+    }
+    ```
+    
+
+交易逻辑配置
+
+- entry/exit_rules:  
+    入口/出口规则：
+    
+    - 仅生成多头买入信号，基于回归模型预测（如 &-s_close > 阈值）。
+        
+    - 退出信号基于动态止盈（如 &-roi_0_pred）或止损（如 &-stoploss_pred）。
+        
+    - 示例：
+        
+        json
+        
+        ```json
+        {
+          "entry_rules": {
+            "conditions": [
+              {"&-action": "buy"},  // 回归模型预测买入
+              {"&-buy_rsi_pred": "> 30"}  // RSI 阈值
+            ]
+          },
+          "exit_rules": {
+            "conditions": [
+              {"&-roi_0_pred": "reached"},  // 达到止盈
+              {"&-stoploss_pred": "triggered"}  // 触发止损
+            ]
+          }
+        }
+        ```
+        
+
+风险管理配置
+
+- stake_amount:  
+    投注金额：
+    
+    - 设置单笔交易风险，如账户余额的 2%。
+        
+    - 示例：
+        
+        json
+        
+        ```json
+        {
+          "stake_amount": "dynamic",
+          "max_open_trades": 3,
+          "position_size": "2% of balance"
+        }
+        ```
+        
+- stoploss:  
+    止损：
+    
+    - 配置动态止损，结合 trailing stop。
+        
+    - 示例：
+        
+        json
+        
+        ```json
+        {
+          "stoploss": {
+            "type": "dynamic",
+            "value": "&-stoploss_pred",
+            "trailing_stop": true,
+            "trailing_stop_positive": 0.01
+          }
+        }
+        ```
+        
+
+日志与监控配置
+
+- logging:  
+    记录日志：
+    
+    json
+    
+    ```json
+    {
+      "logging": {
+        "level": "DEBUG",
+        "logfile": "freqtrade_okx.log",
+        "freqai_log": {
+          "feature_generation": true,
+          "model_performance": true,
+          "data_source_status": true  // 记录 OKX API 状态
+        }
+      }
+    }
+    ```
+    
+
+---
+
+总结与核心原则
+
+本策略围绕以下核心目标展开，全面适配 OKX 数据、只做多和回归模型要求：
+
+1. 标签列的明确与初始化：基于 OKX 数据，通过 set_freqai_targets 和 feature_engineering_expand_all 初始化标签，涵盖目标变量、特征列、预测结果列和元数据列。
+    
+2. 程序鲁棒性增强：通过 OKX 数据校验、日志记录、异常处理和依赖管理，确保策略稳定运行。
+    
+3. 防止未来数据泄露：通过时序约束、OKX 数据划分、审计日志和自动化检测，确保回测和实盘不使用未来数据。
+    
+4. 风险与资金管理：通过动态止损、仓位管理和极端事件处理，平衡多头盈利与风险。
+    
+5. 数据源可靠性：确保 OKX 数据通过 CCXT 获取可靠，符合合规性要求。
+    
+6. 用户交互与监控：提供实时监控和报警机制，提升用户体验。
+    
+7. 策略可扩展性与可维护性：通过模块化设计、配置驱动和接口抽象，适应长期开发需求。
+    
+8. 循序渐进优化：从简单回归模型到复杂特征，逐步实现 OKX 市场稳定盈利。
+    
+
+强调的重点
+
+- 数据一致性：所有数据来自 OKX，通过 CCXT 确保可靠性和合规性。
+    
+- 只做多逻辑：严格限制交易信号为买入和持有，避免空头操作。
+    
+- 回归模型优化：优先回归模型的预测准确性和稳定性，定期评估新模型。
+    
+- 鲁棒性：在 OKX 数据处理、特征生成和信号生成中加入异常处理。
+    
+- 全面评估：通过 MSE、夏普比率、最大回撤等指标，验证 OKX 市场盈利能力。
+    
+
+通过以上设计和配置，本策略能够在 OKX 市场环境中实现稳定多头盈利，同时保持高度的灵活性和鲁棒性。未来可根据 OKX 市场变化或新需求，进一步扩展特征或优化回归模型。
+
+---

nohup.out

@@ -0,0 +1,238 @@
+mv: cannot stat './user_data/backtest_results/*': No such file or directory
+unzip:  cannot find or open *.zip, *.zip.zip or *.zip.ZIP.
+
+No zipfiles found.
+rm: cannot remove '*.zip': No such file or directory
+rm: cannot remove '*.feather': No such file or directory
+/home/ubuntu/freqtrade
+mv: cannot stat './user_data/backtest_results/*': No such file or directory
+unzip:  cannot find or open *.zip, *.zip.zip or *.zip.ZIP.
+
+No zipfiles found.
+rm: cannot remove '*.zip': No such file or directory
+rm: cannot remove '*.feather': No such file or directory
+/home/ubuntu/freqtrade
+mv: cannot stat './user_data/backtest_results/*': No such file or directory
+unzip:  cannot find or open *.zip, *.zip.zip or *.zip.ZIP.
+
+No zipfiles found.
+rm: cannot remove '*.zip': No such file or directory
+rm: cannot remove '*.feather': No such file or directory
+/home/ubuntu/freqtrade
+mv: cannot stat './user_data/backtest_results/*': No such file or directory
+unzip:  cannot find or open *.zip, *.zip.zip or *.zip.ZIP.
+
+No zipfiles found.
+rm: cannot remove '*.zip': No such file or directory
+rm: cannot remove '*.feather': No such file or directory
+/home/ubuntu/freqtrade
+mv: cannot stat './user_data/backtest_results/*': No such file or directory
+unzip:  cannot find or open *.zip, *.zip.zip or *.zip.ZIP.
+
+No zipfiles found.
+rm: cannot remove '*.zip': No such file or directory
+rm: cannot remove '*.feather': No such file or directory
+/home/ubuntu/freqtrade
+mv: cannot stat './user_data/backtest_results/*': No such file or directory
+unzip:  cannot find or open *.zip, *.zip.zip or *.zip.ZIP.
+
+No zipfiles found.
+rm: cannot remove '*.zip': No such file or directory
+rm: cannot remove '*.feather': No such file or directory
+/home/ubuntu/freqtrade
+mv: cannot stat './user_data/backtest_results/*': No such file or directory
+unzip:  cannot find or open *.zip, *.zip.zip or *.zip.ZIP.
+
+No zipfiles found.
+rm: cannot remove '*.zip': No such file or directory
+rm: cannot remove '*.feather': No such file or directory
+/home/ubuntu/freqtrade
+mv: cannot stat './user_data/backtest_results/*': No such file or directory
+unzip:  cannot find or open *.zip, *.zip.zip or *.zip.ZIP.
+
+No zipfiles found.
+rm: cannot remove '*.zip': No such file or directory
+rm: cannot remove '*.feather': No such file or directory
+/home/ubuntu/freqtrade
+mv: cannot stat './user_data/backtest_results/*': No such file or directory
+unzip:  cannot find or open *.zip, *.zip.zip or *.zip.ZIP.
+
+No zipfiles found.
+rm: cannot remove '*.zip': No such file or directory
+rm: cannot remove '*.feather': No such file or directory
+/home/ubuntu/freqtrade
+mv: cannot stat './user_data/backtest_results/*': No such file or directory
+unzip:  cannot find or open *.zip, *.zip.zip or *.zip.ZIP.
+
+No zipfiles found.
+rm: cannot remove '*.zip': No such file or directory
+rm: cannot remove '*.feather': No such file or directory
+/home/ubuntu/freqtrade
+mv: cannot stat './user_data/backtest_results/*': No such file or directory
+unzip:  cannot find or open *.zip, *.zip.zip or *.zip.ZIP.
+
+No zipfiles found.
+rm: cannot remove '*.zip': No such file or directory
+rm: cannot remove '*.feather': No such file or directory
+/home/ubuntu/freqtrade
+mv: cannot stat './user_data/backtest_results/*': No such file or directory
+unzip:  cannot find or open *.zip, *.zip.zip or *.zip.ZIP.
+
+No zipfiles found.
+rm: cannot remove '*.zip': No such file or directory
+rm: cannot remove '*.feather': No such file or directory
+/home/ubuntu/freqtrade
+mv: cannot stat './user_data/backtest_results/*': No such file or directory
+unzip:  cannot find or open *.zip, *.zip.zip or *.zip.ZIP.
+
+No zipfiles found.
+rm: cannot remove '*.zip': No such file or directory
+rm: cannot remove '*.feather': No such file or directory
+/home/ubuntu/freqtrade
+mv: cannot stat './user_data/backtest_results/*': No such file or directory
+unzip:  cannot find or open *.zip, *.zip.zip or *.zip.ZIP.
+
+No zipfiles found.
+rm: cannot remove '*.zip': No such file or directory
+rm: cannot remove '*.feather': No such file or directory
+/home/ubuntu/freqtrade
+mv: cannot stat './user_data/backtest_results/*': No such file or directory
+unzip:  cannot find or open *.zip, *.zip.zip or *.zip.ZIP.
+
+No zipfiles found.
+rm: cannot remove '*.zip': No such file or directory
+rm: cannot remove '*.feather': No such file or directory
+/home/ubuntu/freqtrade
+mv: cannot stat './user_data/backtest_results/*': No such file or directory
+unzip:  cannot find or open *.zip, *.zip.zip or *.zip.ZIP.
+
+No zipfiles found.
+rm: cannot remove '*.zip': No such file or directory
+rm: cannot remove '*.feather': No such file or directory
+/home/ubuntu/freqtrade
+mv: cannot stat './user_data/backtest_results/*': No such file or directory
+unzip:  cannot find or open *.zip, *.zip.zip or *.zip.ZIP.
+
+No zipfiles found.
+rm: cannot remove '*.zip': No such file or directory
+rm: cannot remove '*.feather': No such file or directory
+/home/ubuntu/freqtrade
+mv: cannot stat './user_data/backtest_results/*': No such file or directory
+unzip:  cannot find or open *.zip, *.zip.zip or *.zip.ZIP.
+
+No zipfiles found.
+rm: cannot remove '*.zip': No such file or directory
+rm: cannot remove '*.feather': No such file or directory
+/home/ubuntu/freqtrade
+mv: cannot stat './user_data/backtest_results/*': No such file or directory
+unzip:  cannot find or open *.zip, *.zip.zip or *.zip.ZIP.
+
+No zipfiles found.
+rm: cannot remove '*.zip': No such file or directory
+rm: cannot remove '*.feather': No such file or directory
+/home/ubuntu/freqtrade
+mv: cannot stat './user_data/backtest_results/*': No such file or directory
+unzip:  cannot find or open *.zip, *.zip.zip or *.zip.ZIP.
+
+No zipfiles found.
+rm: cannot remove '*.zip': No such file or directory
+rm: cannot remove '*.feather': No such file or directory
+/home/ubuntu/freqtrade
+mv: cannot stat './user_data/backtest_results/*': No such file or directory
+unzip:  cannot find or open *.zip, *.zip.zip or *.zip.ZIP.
+
+No zipfiles found.
+rm: cannot remove '*.zip': No such file or directory
+rm: cannot remove '*.feather': No such file or directory
+/home/ubuntu/freqtrade
+Archive:  backtest-result-2025-05-06_01-08-03.zip
+  inflating: backtest-result-2025-05-06_01-08-03.json  
+  inflating: backtest-result-2025-05-06_01-08-03_config.json  
+  inflating: backtest-result-2025-05-06_01-08-03_OKXRegressionStrategy.py  
+  inflating: backtest-result-2025-05-06_01-08-03_market_change.feather  
+/home/ubuntu/freqtrade
+Archive:  backtest-result-2025-05-06_01-28-09.zip
+  inflating: backtest-result-2025-05-06_01-28-09.json  
+  inflating: backtest-result-2025-05-06_01-28-09_config.json  
+  inflating: backtest-result-2025-05-06_01-28-09_OKXRegressionStrategy.py  
+  inflating: backtest-result-2025-05-06_01-28-09_market_change.feather  
+/home/ubuntu/freqtrade
+Archive:  backtest-result-2025-05-06_01-53-53.zip
+  inflating: backtest-result-2025-05-06_01-53-53.json  
+  inflating: backtest-result-2025-05-06_01-53-53_config.json  
+  inflating: backtest-result-2025-05-06_01-53-53_OKXRegressionStrategy.py  
+  inflating: backtest-result-2025-05-06_01-53-53_market_change.feather  
+/home/ubuntu/freqtrade
+Archive:  backtest-result-2025-05-06_02-15-16.zip
+  inflating: backtest-result-2025-05-06_02-15-16.json  
+  inflating: backtest-result-2025-05-06_02-15-16_config.json  
+  inflating: backtest-result-2025-05-06_02-15-16_OKXRegressionStrategy.py  
+  inflating: backtest-result-2025-05-06_02-15-16_market_change.feather  
+/home/ubuntu/freqtrade
+mv: cannot stat './user_data/backtest_results/*': No such file or directory
+unzip:  cannot find or open *.zip, *.zip.zip or *.zip.ZIP.
+
+No zipfiles found.
+rm: cannot remove '*.zip': No such file or directory
+rm: cannot remove '*.feather': No such file or directory
+/home/ubuntu/freqtrade
+Archive:  backtest-result-2025-05-06_02-50-43.zip
+  inflating: backtest-result-2025-05-06_02-50-43.json  
+  inflating: backtest-result-2025-05-06_02-50-43_config.json  
+  inflating: backtest-result-2025-05-06_02-50-43_OKXRegressionStrategy.py  
+  inflating: backtest-result-2025-05-06_02-50-43_market_change.feather  
+/home/ubuntu/freqtrade
+Archive:  backtest-result-2025-05-06_03-05-54.zip
+  inflating: backtest-result-2025-05-06_03-05-54.json  
+  inflating: backtest-result-2025-05-06_03-05-54_config.json  
+  inflating: backtest-result-2025-05-06_03-05-54_OKXRegressionStrategy.py  
+  inflating: backtest-result-2025-05-06_03-05-54_market_change.feather  
+/home/ubuntu/freqtrade
+./run.sh: line 33: result: Is a directory
+mv: cannot stat './user_data/backtest_results/*': No such file or directory
+unzip:  cannot find or open *.zip, *.zip.zip or *.zip.ZIP.
+
+No zipfiles found.
+rm: cannot remove '*.zip': No such file or directory
+rm: cannot remove '*.feather': No such file or directory
+/home/ubuntu/freqtrade
+mv: cannot stat './user_data/backtest_results/*': No such file or directory
+unzip:  cannot find or open *.zip, *.zip.zip or *.zip.ZIP.
+
+No zipfiles found.
+rm: cannot remove '*.zip': No such file or directory
+rm: cannot remove '*.feather': No such file or directory
+/home/ubuntu/freqtrade
+mv: cannot stat './user_data/backtest_results/*': No such file or directory
+unzip:  cannot find or open *.zip, *.zip.zip or *.zip.ZIP.
+
+No zipfiles found.
+rm: cannot remove '*.zip': No such file or directory
+rm: cannot remove '*.feather': No such file or directory
+/home/ubuntu/freqtrade
+mv: cannot stat './user_data/backtest_results/*': No such file or directory
+unzip:  cannot find or open *.zip, *.zip.zip or *.zip.ZIP.
+
+No zipfiles found.
+rm: cannot remove '*.zip': No such file or directory
+rm: cannot remove '*.feather': No such file or directory
+/home/ubuntu/freqtrade
+mv: cannot stat './user_data/backtest_results/*': No such file or directory
+unzip:  cannot find or open *.zip, *.zip.zip or *.zip.ZIP.
+
+No zipfiles found.
+rm: cannot remove '*.zip': No such file or directory
+rm: cannot remove '*.feather': No such file or directory
+/home/ubuntu/freqtrade
+Archive:  backtest-result-2025-05-06_03-47-33.zip
+  inflating: backtest-result-2025-05-06_03-47-33.json  
+  inflating: backtest-result-2025-05-06_03-47-33_config.json  
+  inflating: backtest-result-2025-05-06_03-47-33_OKXRegressionStrategy.py  
+  inflating: backtest-result-2025-05-06_03-47-33_market_change.feather  
+/home/ubuntu/freqtrade
+Archive:  backtest-result-2025-05-06_04-07-15.zip
+  inflating: backtest-result-2025-05-06_04-07-15.json  
+  inflating: backtest-result-2025-05-06_04-07-15_config.json  
+  inflating: backtest-result-2025-05-06_04-07-15_OKXRegressionStrategy.py  
+  inflating: backtest-result-2025-05-06_04-07-15_market_change.feather  
+/home/ubuntu/freqtrade

run.sh

@@ -0,0 +1,30 @@
+#!/bin/bash
+
+rm -rf user_data/models/*
+rm -rf ./freqtrade/user_data/data/backtest_results/*
+
+docker-compose run --rm freqtrade >output.log 2>&1
+sed -i 's/\x1B\[[0-9;]*m//g' output.log
+python3 filter.py
+
+rm ./result/* -fr
+mv ./user_data/backtest_results/* ./result/
+
+cd ./result
+# 查找当前目录下的所有 zip 文件
+zip_files=(*.zip)
+
+# 检查是否只有一个 zip 文件
+if [ ${#zip_files[@]} -eq 1 ]; then
+  # 解压缩该 zip 文件到当前目录
+  unzip "${zip_files[0]}"
+  rm *.zip
+  rm *.feather
+else
+  echo "当前目录下没有 zip 文件或者有多个 zip 文件，无法操作。"
+fi
+
+cd -
+sed -i 's/\x1B\[[0-9;]*m//g' output.log
+python3 filter.py
+cp output_filted.log result/ -f

run.sh_disabled

@@ -0,0 +1,30 @@
+#!/bin/bash
+
+rm -rf user_data/models/*
+rm -rf ./freqtrade/user_data/data/backtest_results/*
+
+docker-compose run --rm freqtrade >output.log 2>&1
+sed -i 's/\x1B\[[0-9;]*m//g' output.log
+python3 filter.py
+
+rm ./result/* -fr
+mv ./user_data/backtest_results/* ./result/
+
+cd ./result
+# 查找当前目录下的所有 zip 文件
+zip_files=(*.zip)
+
+# 检查是否只有一个 zip 文件
+if [ ${#zip_files[@]} -eq 1 ]; then
+  # 解压缩该 zip 文件到当前目录
+  unzip "${zip_files[0]}"
+  rm *.zip
+  rm *.feather
+else
+  echo "当前目录下没有 zip 文件或者有多个 zip 文件，无法操作。"
+fi
+
+cd -
+sed -i 's/\x1B\[[0-9;]*m//g' output.log
+python3 filter.py
+cp output_filted.log result/ -f

tests/config_test_comments.json

@@ -0,0 +1,129 @@
+{
+    /* Single-line C-style comment */
+    "max_open_trades": 3,
+    /*
+     * Multi-line C-style comment
+     */
+    "stake_currency": "BTC",
+    "stake_amount": 0.05,
+    "fiat_display_currency": "USD", // C++-style comment
+    "amount_reserve_percent": 0.05, // And more, tabs before this comment
+    "dry_run": false,
+    "timeframe": "5m",
+    "trailing_stop": false,
+    "trailing_stop_positive": 0.005,
+    "trailing_stop_positive_offset": 0.0051,
+    "trailing_only_offset_is_reached": false,
+    "minimal_roi": {
+        "40": 0.0,
+        "30": 0.01,
+        "20": 0.02,
+        "0": 0.04
+    },
+    "stoploss": -0.10,
+    "unfilledtimeout": {
+        "buy": 10,
+        "sell": 30, // Trailing comma should also be accepted now
+    },
+    "bid_strategy": {
+        "use_order_book": false,
+        "ask_last_balance": 0.0,
+        "order_book_top": 1,
+        "check_depth_of_market": {
+            "enabled": false,
+            "bids_to_ask_delta": 1
+        }
+    },
+    "ask_strategy": {
+        "use_order_book": false,
+        "order_book_min": 1,
+        "order_book_max": 9
+    },
+    "order_types": {
+        "buy": "limit",
+        "sell": "limit",
+        "stoploss": "market",
+        "stoploss_on_exchange": false,
+        "stoploss_on_exchange_interval": 60
+    },
+    "order_time_in_force": {
+        "buy": "gtc",
+        "sell": "gtc"
+    },
+    "pairlist": {
+        "method": "VolumePairList",
+        "config": {
+            "number_assets": 20,
+            "sort_key": "quoteVolume",
+            "precision_filter": false
+        }
+    },
+    "exchange": {
+        "name": "binance",
+        "sandbox": false,
+        "key": "your_exchange_key",
+        "secret": "your_exchange_secret",
+        "password": "",
+        "ccxt_config": {
+            "enableRateLimit": true
+        },
+        "ccxt_async_config": {
+            "enableRateLimit": false,
+            "rateLimit": 500,
+            "aiohttp_trust_env": false
+        },
+        "pair_whitelist": [
+            "ETH/BTC",
+            "LTC/BTC",
+            "ETC/BTC",
+            "DASH/BTC",
+            "ZEC/BTC",
+            "XLM/BTC",
+            "NXT/BTC",
+            "TRX/BTC",
+            "ADA/BTC",
+            "XMR/BTC"
+        ],
+        "pair_blacklist": [
+            "DOGE/BTC"
+        ],
+        "outdated_offset": 5,
+        "markets_refresh_interval": 60
+    },
+    "edge": {
+        "enabled": false,
+        "process_throttle_secs": 3600,
+        "calculate_since_number_of_days": 7,
+        "allowed_risk": 0.01,
+        "stoploss_range_min": -0.01,
+        "stoploss_range_max": -0.1,
+        "stoploss_range_step": -0.01,
+        "minimum_winrate": 0.60,
+        "minimum_expectancy": 0.20,
+        "min_trade_number": 10,
+        "max_trade_duration_minute": 1440,
+        "remove_pumps": false
+    },
+    "telegram": {
+        // We can now comment out some settings
+        //        "enabled": true,
+        "enabled": false,
+        "token": "your_telegram_token",
+        "chat_id": "your_telegram_chat_id"
+    },
+    "api_server": {
+        "enabled": false,
+        "listen_ip_address": "127.0.0.1",
+        "listen_port": 8080,
+        "username": "freqtrader",
+        "password": "SuperSecurePassword"
+    },
+    "db_url": "sqlite:///tradesv3.sqlite",
+    "initial_state": "running",
+    "forcebuy_enable": false,
+    "internals": {
+        "process_throttle_secs": 5
+    },
+    "strategy": "DefaultStrategy",
+    "strategy_path": "user_data/strategies/"
+}

tools/analytic.py

@@ -0,0 +1,74 @@
+import pandas as pd
+
+# 加载交易记录
+df = pd.read_csv('../result/backtest_trades.csv')
+
+# 转换日期格式
+df['open_date'] = pd.to_datetime(df['open_date'])
+df['close_date'] = pd.to_datetime(df['close_date'])
+
+# 计算持仓天数
+df['holding_days'] = (df['close_date'] - df['open_date']).dt.total_seconds() / (60 * 60 * 24)
+
+# 按币种分组
+grouped_by_pair = df.groupby('pair').agg(
+    total_trades=('profit_abs', 'size'),
+    avg_profit_ratio=('profit_ratio', 'mean'),
+    total_profit_abs=('profit_abs', 'sum'),
+    win_rate=('profit_ratio', lambda x: (x > 0).mean()),
+    avg_duration=('trade_duration', 'mean')
+)
+
+print(grouped_by_pair)
+
+#按退出原因分组
+grouped_by_exit = df.groupby('exit_reason').agg(
+    total_trades=('profit_abs', 'size'),
+    avg_profit_ratio=('profit_ratio', 'mean'),
+    total_profit_abs=('profit_abs', 'sum'),
+    win_rate=('profit_ratio', lambda x: (x > 0).mean())
+)
+
+print(grouped_by_exit)
+#按月份分组统计
+df['open_date_naive'] = df['open_date'].dt.tz_localize(None)
+df['month'] = df['open_date_naive'].dt.to_period('M')
+
+
+grouped_by_month = df.groupby('month').agg(
+    total_trades=('profit_abs', 'size'),
+    avg_profit_ratio=('profit_ratio', 'mean'),
+    total_profit_abs=('profit_abs', 'sum'),
+    win_rate=('profit_ratio', lambda x: (x > 0).mean())
+)
+
+print(grouped_by_month)
+# 按盈利区间分组统计
+bins = [-float('inf'), -0.05, -0.02, 0, 0.02, 0.05, float('inf')]
+labels = ['<-5%', '-5%~-2%', '-2%~0%', '0%~2%', '2%~5%', '>5%']
+
+df['profit_group'] = pd.cut(df['profit_ratio'], bins=bins, labels=labels)
+
+grouped_by_profit = df.groupby('profit_group', observed=True).agg(
+    count=('profit_abs', 'size'),
+    avg_profit=('profit_ratio', 'mean'),
+    total_profit=('profit_abs', 'sum')
+)
+
+print(grouped_by_profit)
+# 分组为短中长线
+df['duration_group'] = pd.cut(df['trade_duration'], 
+                              bins=[0, 60, 360, 1440, 100000], 
+                              labels=['<1小时', '1~6小时', '6小时~1天', '>1天'])
+
+grouped_by_duration = df.groupby('duration_group', observed=True).agg(
+    count=('profit_abs', 'size'),
+    avg_profit=('profit_ratio', 'mean'),
+    win_rate=('profit_ratio', lambda x: (x > 0).mean()),
+    total_profit=('profit_abs', 'sum')
+)
+
+print(grouped_by_duration)
+
+
+

tools/analytic.sh

@@ -0,0 +1,4 @@
+cd ../
+source .venv/bin/activate
+cd -
+python analytic.py

tools/backtest.sh

@@ -0,0 +1,106 @@
+#!/bin/bash
+
+# 检查 .env 文件
+if [ ! -f ".env" ]; then
+  echo "⚠️ 本地缺少 .env 文件，请创建并配置。示例内容如下："
+  echo ""
+  echo "STRATEGY_NAME=TheForceV7"
+  echo "CONFIG_FILE=basic.json"
+  echo "TEST_BRANCH=theforce-noai-test"
+  echo "DRYRUN_BRANCH=theforce-noai-dryrun"
+  echo ""
+  exit 1
+fi
+
+# 加载 .env 文件中的变量
+export $(grep -v '^#' .env | xargs)
+# 设置默认值
+STRATEGY_NAME=${STRATEGY_NAME:-TheForceV7}
+CONFIG_FILE=${CONFIG_FILE:-basic.json}
+
+echo "Using strategy: $STRATEGY_NAME"
+echo "Using config: $CONFIG_FILE"
+echo "Using testBranch: $TEST_BRANCH"
+# Parse command line arguments
+START_DATE=${1:-$(date -d "2 days ago" +"%Y%m%d")}
+END_DATE=${2:-$(date -d "tomorrow" +"%Y%m%d")}
+PARAMS_NAME=$(echo "$STRATEGY_NAME" | tr '[:upper:]' '[:lower:]')
+
+# Parse command line arguments
+START_DATE_RAW=${1:-$(date -d "2 days ago" +"%Y-%m-%d %H:%M:%S")}
+END_DATE_RAW=${2:-$(date -d "tomorrow" +"%Y-%m-%d %H:%M:%S")}
+
+# 判断是否包含空格，如果是，则按完整时间解析并转为时间戳
+if [[ "$START_DATE_RAW" == *" "* ]]; then
+  START_DATE=$(date -d "$START_DATE_RAW" +"%s")
+else
+  # 兼容旧格式：20250616 -> 转换为 2025-06-16 00:00:00 再转为时间戳
+  START_DATE=$(date -d "${START_DATE_RAW:0:4}-${START_DATE_RAW:4:2}-${START_DATE_RAW:6:2} 00:00:00" +"%s")
+fi
+
+if [[ "$END_DATE_RAW" == *" "* ]]; then
+  END_DATE=$(date -d "$END_DATE_RAW" +"%s")
+else
+  END_DATE=$(date -d "${END_DATE_RAW:0:4}-${END_DATE_RAW:4:2}-${END_DATE_RAW:6:2} 00:00:00" +"%s")
+fi
+
+cd ../
+source .venv/bin/activate
+rm -rf user_data/models/*
+rm -rf ./freqtrade/user_data/data/backtest_results/*
+rm -fr ./user_data/dryrun_results/*
+rm result/*
+
+hyperopt_config="${STRATEGY_NAME%.py}.json"
+#docker-compose -f docker-compose_backtest.yml run --rm freqtrade >output.log 2>&1
+
+echo "docker-compose run --rm freqtrade backtesting \
+  --logfile /freqtrade/user_data/logs/freqtrade.log \
+  --freqaimodel LightGBMRegressorMultiTarget \
+  --config /freqtrade/config_examples/$CONFIG_FILE \
+  --strategy-path /freqtrade/templates \
+  --strategy $STRATEGY_NAME \
+  --timerange $START_DATE-$END_DATE \
+  --fee 0.0008 \
+  --cache none >output.log"
+#LightGBMRegressorMultiTarget
+docker-compose run --rm freqtrade backtesting \
+  --logfile /freqtrade/user_data/logs/freqtrade.log \
+  --freqaimodel LightGBMRegressorMultiTarget \
+  --config /freqtrade/config_examples/$CONFIG_FILE \
+  --strategy-path /freqtrade/templates \
+  --strategy $STRATEGY_NAME \
+  --timerange $START_DATE-$END_DATE \
+  --fee 0.0008 \
+  --breakdown day \
+  --cache none >output.log 2>&1
+sed -i 's/\x1B\[[0-9;]*m//g' output.log
+
+#python3 tools/filter.py
+
+rm ./result/*.json -fr
+rm ./result/*.py -fr
+mv ./user_data/backtest_results/* ./result/
+
+cd ./result
+# 查找当前目录下的所有 zip 文件
+zip_files=(*.zip)
+
+# 检查是否只有一个 zip 文件
+if [ ${#zip_files[@]} -eq 1 ]; then
+  # 解压缩该 zip 文件到当前目录
+  unzip "${zip_files[0]}"
+  rm *.zip
+  rm *.feather
+else
+  echo "当前目录下没有 zip 文件或者有多个 zip 文件，无法操作。"
+fi
+
+cd -
+sed -i 's/\x1B\[[0-9;]*m//g' output.log
+#python3 ../filter.py
+cp output.log result/ -f
+cd tools/
+python tradestocsv.py
+python analytic.py >../result/analytic.log
+cd ../

tools/download.sh

@@ -0,0 +1,99 @@
+#!/bin/bash
+
+# Function to extract the value of a parameter
+get_param_value() {
+  local param="$1"
+  shift
+  while [[ $# -gt 0 ]]; do
+    case "$1" in
+    $param=*)
+      echo "${1#*=}"
+      return
+      ;;
+    $param)
+      # Check if the next argument exists and does not start with a dash
+      if [[ -n "$2" && "$2" != -* ]]; then
+        echo "$2"
+        return
+      else
+        echo "Error: Missing value for parameter $param" >&2
+        exit 1
+      fi
+      ;;
+    esac
+    shift
+  done
+  echo "Error: Parameter $param not found" >&2
+  exit 1
+}
+
+# Function to extract comma-separated values
+get_csv_param_value() {
+  local param="$1"
+  shift
+  while [[ $# -gt 0 ]]; do
+    case "$1" in
+    $param=*)
+      echo "${1#*=}"
+      return
+      ;;
+    $param)
+      # Check if the next argument exists and does not start with a dash
+      if [[ -n "$2" && "$2" != -* ]]; then
+        echo "$2"
+        return
+      else
+        echo "Error: Missing value for parameter $param" >&2
+        exit 1
+      fi
+      ;;
+    esac
+    shift
+  done
+  echo ""
+  return
+}
+
+if [[ "$@" == *"--timerange"* ]] && [[ "$@" == *"--days"* ]]; then
+  echo "Error: Both --timerange and --days cannot be provided."
+  exit 1
+fi
+
+# Get timerange or days from parameters
+timerange=""
+days=""
+if [[ "$@" == *"--timerange"* ]]; then
+  timerange=$(get_param_value "--timerange" "$@")
+elif [[ "$@" == *"--days"* ]]; then
+  days=$(get_param_value "--days" "$@")
+fi
+
+# Get pairs and timeframe from parameters or use defaults
+pairs=$(get_csv_param_value "--pairs" "$@")
+timeframe=$(get_csv_param_value "--timeframe" "$@")
+
+# Use default values if parameters are not provided
+if [[ -z "$pairs" ]]; then
+  pairs="BTC/USDT,OKB/USDT,TON/USDT,DOT/USDT,SOL/USDT,XRP/USDT"
+fi
+
+if [[ -z "$timeframe" ]]; then
+  timeframe="3m,5m,15m,30m,1h,4h,6h,12h,1d"
+fi
+
+# Convert timeframe string to array
+IFS=',' read -r -a timeframe_array <<<"$timeframe"
+timeframe_array_str=$(printf " '%s'" "${timeframe_array[@]}")
+
+# Initialize the base command
+cmd="docker-compose run --rm freqtrade download-data --config /freqtrade/config_examples/basic.json --pairs $pairs --timeframe$timeframe_array_str"
+
+# Add timerange or days if provided
+if [[ -n "$timerange" ]]; then
+  cmd+=" --timerange='$timerange'"
+elif [[ -n "$days" ]]; then
+  cmd+=" --days=$days"
+fi
+
+# Execute the command
+eval "$cmd"

tools/dryrun.sh

@@ -0,0 +1,95 @@
+#!/bin/bash
+
+set -e # 出错立即停止
+# 检查 .env 文件
+if [ ! -f ".env" ]; then
+  echo "⚠️ 本地缺少 .env 文件，请创建并配置。示例内容如下："
+  echo ""
+  echo "STRATEGY_NAME=TheForceV7"
+  echo "CONFIG_FILE=basic.json"
+  echo "TEST_BRANCH=theforce-noai-test"
+  echo "DRYRUN_BRANCH=theforce-noai-dryrun"
+  echo ""
+  exit 1
+fi
+# 加载 .env 文件中的变量（位于项目根目录）
+if [ -f ".env" ]; then
+  export $(grep -v '^#' .env | xargs)
+else
+  echo "⚠️ 缺少 .env 文件，请创建并配置 TEST_BRANCH 和 DRYRUN_BRANCH"
+  exit 1
+fi
+
+# 获取当前分支名
+current_branch=$(git rev-parse --abbrev-ref HEAD)
+
+# 检查当前分支是否匹配 TEST_BRANCH 或 DRYRUN_BRANCH
+if [[ "$current_branch" != "$TEST_BRANCH" && "$current_branch" != "$DRYRUN_BRANCH" ]]; then
+  echo "⚠️ 错误：当前分支 '$current_branch' 不符合环境变量配置要求。"
+  echo "请确保当前分支与 .env 文件中的 TEST_BRANCH 或 DRYRUN_BRANCH 配置一致。"
+  echo "TEST_BRANCH=$TEST_BRANCH"
+  echo "DRYRUN_BRANCH=$DRYRUN_BRANCH"
+  exit 1
+else
+  echo "✅ 当前分支 '$current_branch' 符合环境变量配置要求。"
+fi
+
+# 如果当前分支名称包含 "dryrun"，执行 git reset --hard
+if [[ "$CURRENT_BRANCH" == *"dryrun"* ]]; then
+  echo "当前分支为 '$CURRENT_BRANCH'，正在执行 git reset --hard 以确保代码干净..."
+  git reset --hard
+  if [ $? -ne 0 ]; then
+    echo "⚠️ 执行 git reset --hard 失败，请检查 Git 状态。"
+    exit 1
+  fi
+  echo "Git 工作区已清理。"
+fi
+
+# 加载 .env 文件中的变量
+export $(grep -v '^#' .env | xargs)
+# 设置默认值
+STRATEGY_NAME=${STRATEGY_NAME:-TheForceV7}
+CONFIG_FILE=${CONFIG_FILE:-basic.json}
+PARAMS_NAME=$(echo "$STRATEGY_NAME" | tr '[:upper:]' '[:lower:]')
+
+echo "Using strategy: $STRATEGY_NAME"
+echo "Using config: $CONFIG_FILE"
+echo "Using testBranch: $TEST_BRANCH"
+# Parse command line arguments
+START_DATE=${1:-$(date -d "2 days ago" +"%Y%m%d")}
+END_DATE=${2:-$(date -d "tomorrow" +"%Y%m%d")}
+
+cd ../
+source .venv/bin/activate
+rm -rf user_data/models/*
+rm -rf ./freqtrade/user_data/data/backtest_results/*
+rm -fr ./user_data/dryrun_results/*
+rm result/* || true
+
+hyperopt_config="${STRATEGY_NAME%.py}.json"
+docker rm $(docker ps -aq) -f || true
+
+# 获取当前 Git Commit 的前 8 位
+GIT_COMMIT_SHORT=$(git rev-parse HEAD | cut -c 1-8)
+echo "docker-compose run -d --rm --name freqtrade-dryrun-${GIT_COMMIT_SHORT} -p 8080:8080 freqtrade trade \
+  --logfile /freqtrade/user_data/logs/freqtrade.log \
+  --db-url sqlite:////freqtrade/user_data/tradesv3.sqlite \
+  --dry-run \
+  --freqaimodel LightGBMRegressorMultiTarget \
+  --config /freqtrade/config_examples/$CONFIG_FILE \
+  --config /freqtrade/templates/${PARAMS_NAME}.json \
+  --strategy $STRATEGY_NAME \
+  --config /freqtrade/freqtrade/templates/${STRATEGY_NAME}.json \
+  --fee 0.0008 \
+  --strategy-path /freqtrade/templates"
+
+docker-compose run -d --rm --name freqtrade-dryrun-${GIT_COMMIT_SHORT} -p 8080:8080 freqtrade trade \
+  --logfile /freqtrade/user_data/logs/freqtrade.log \
+  --db-url sqlite:////freqtrade/user_data/tradesv3.sqlite \
+  --dry-run \
+  --freqaimodel LightGBMRegressorMultiTarget \
+  --config /freqtrade/config_examples/$CONFIG_FILE \
+  --config /freqtrade/templates/${PARAMS_NAME}.json \
+  --strategy $STRATEGY_NAME \
+  --fee 0.0008 \
+  --strategy-path /freqtrade/templates

tools/dryrunloop.service

@@ -0,0 +1,16 @@
+[Unit]
+Description=Run dryrunloop.sh script daily
+After=network.target
+
+[Service]
+Type=oneshot
+ExecStart=/bin/bash /home/ubuntu/freqtrade/tools/dryrunloop.sh
+WorkingDirectory=/home/ubuntu/freqtrade/tools
+User=ubuntu
+Group=ubuntu
+
+# Ensure the script exits cleanly
+SuccessExitStatus=0
+
+[Install]
+WantedBy=multi-user.target

tools/dryrunloop.sh

@@ -0,0 +1,10 @@
+#!/bin/bash
+
+# 动态获取昨天和明天的日期
+START_DATE=$(date -d "yesterday" +"%Y%m%d")
+END_DATE=$(date -d "tomorrow" +"%Y%m%d")
+
+./download.sh --timerange $START_DATE-$END_DATE
+# 执行 hyperopt 和 dryrun
+./hyperopt.sh $START_DATE $END_DATE
+./dryrun.sh

tools/dryrunloop.timer

@@ -0,0 +1,9 @@
+[Unit]
+Description=Run dryrunloop.sh daily at 8:05 AM
+
+[Timer]
+OnCalendar=*-*-* 08:05:00
+Persistent=true
+
+[Install]
+WantedBy=timers.target

tools/exchanges.py

@@ -0,0 +1,45 @@
+import pandas as pd
+
+# 加载回测结果
+df = pd.read_csv('../result/backtest_trades.csv')
+
+# 转换时间为 datetime 类型
+df['open_date'] = pd.to_datetime(df['open_date'])
+df['close_date'] = pd.to_datetime(df['close_date'])
+
+# 按币种分组
+grouped = df.groupby('pair')
+
+# 存储最终结果
+results = []
+
+for pair, group in grouped:
+    # 按照开仓时间排序
+    group = group.sort_values(by='open_date')
+
+    for _, trade in group.iterrows():
+        entry_time = trade['open_date']
+        exit_time = trade['close_date']
+        entry_price = trade['open_rate']
+        exit_price = trade['close_rate']
+        profit_abs = trade['profit_abs']
+        profit_ratio = trade['profit_ratio']
+
+        results.append({
+            '币种': pair,
+            '入场时间': entry_time,
+            '入场价格': entry_price,
+            '出场时间': exit_time,
+            '出场价格': exit_price,
+            '盈利金额': profit_abs,
+            '盈利比例': f"{profit_ratio * 100:.2f}%"
+        })
+
+# 输出为 DataFrame 并保存到 CSV
+trade_log = pd.DataFrame(results)
+trade_log.to_csv('../result/trade_log_detailed.csv', index=False)
+
+print("✅ 已生成详细交易日志：../result/trade_log_detailed.csv")
+print(trade_log)
+
+

tools/exchanges.sh

@@ -0,0 +1,4 @@
+cd ../
+source .venv/bin/activate
+cd -
+python exchanges.py

tools/filter.py

@@ -0,0 +1,11 @@
+import re
+
+pattern = re.compile(r'validation_0')
+
+with open('output.log', 'r') as input_file, open('output_filted.log', 'w') as output_file:
+    for line in input_file:
+        if "validation_0" not in line:
+            output_file.write(line)
+        else:
+            if "99" in line[:5]:
+                output_file.write(line)

tools/gitpush.sh

@@ -0,0 +1,48 @@
+#!/bin/bash
+
+#!/bin/bash
+
+# 默认提交信息
+commit_msg="update"
+
+# 解析命令行参数 -m 指定提交信息
+while getopts m: option; do
+  case "${option}" in
+  m) commit_msg=${OPTARG} ;;
+  esac
+done
+
+cd ..
+
+# 检查 result/ 是否已经在 .gitignore 中
+IGNORED=$(grep -c "^result/" .gitignore)
+
+# 创建标志文件路径
+LOCK_FILE=".gitignore.lock"
+
+# 如果 result/ 不在 .gitignore 中，则临时加入，并记录
+if [ "$IGNORED" -eq 0 ]; then
+  echo "result/" >>.gitignore
+  echo "已临时将 result/ 添加到 .gitignore"
+  touch $LOCK_FILE
+else
+  echo "result/ 已经在 .gitignore 中，跳过处理"
+fi
+
+rm output*.log
+# 停止当前可能存在的交易进程（可选）
+# pkill -f "freqtrade trade" || true
+
+# 执行 Git 操作
+git add .
+git commit --no-verify -m "$commit_msg"
+git push
+
+# 如果之前临时加入了 result/ 到 .gitignore，则现在恢复
+if [ -f "$LOCK_FILE" ]; then
+  sed -i '/^result\//d' .gitignore
+  rm -f $LOCK_FILE
+  echo "已恢复 .gitignore，result/ 重新纳入版本控制"
+fi
+
+echo "Git push 完成"

tools/hyperopt.sh

@@ -0,0 +1,113 @@
+#!/bin/bash
+
+# 检查 .env 文件
+if [ ! -f ".env" ]; then
+  echo "⚠️ 本地缺少 .env 文件，请创建并配置。示例内容如下："
+  echo ""
+  echo "STRATEGY_NAME=FreqaiPrimer"
+  echo "CONFIG_FILE=FreqaiPrimer.json"
+  echo "TEST_BRANCH=freqaiprimer-deviation-test"
+  echo "DRYRUN_BRANCH=freqaiprimer-deviation-dryrun"
+  echo ""
+  exit 1
+fi
+
+# 加载 .env 文件中的变量
+export $(grep -v '^#' .env | xargs)
+# 设置默认值
+STRATEGY_NAME=${STRATEGY_NAME:-FreqaiPrimer}
+CONFIG_FILE=${CONFIG_FILE:-FreqaiPrimer.json}
+PARAMS_NAME=$(echo "$STRATEGY_NAME" | tr '[:upper:]' '[:lower:]')
+
+echo "Using strategy: $STRATEGY_NAME"
+echo "Using config: $CONFIG_FILE"
+echo "Using testBranch: $TEST_BRANCH"
+# Parse command line arguments
+START_DATE=${1:-$(date -d "2 days ago" +"%Y%m%d")}
+END_DATE=${2:-$(date -d "tomorrow" +"%Y%m%d")}
+# Parse command line arguments
+START_DATE_RAW=${1:-$(date -d "2 days ago" +"%Y-%m-%d %H:%M:%S")}
+END_DATE_RAW=${2:-$(date -d "tomorrow" +"%Y-%m-%d %H:%M:%S")}
+
+# Parse command line arguments
+START_DATE_RAW=${1:-$(date -d "2 days ago" +"%Y-%m-%d 00:00")}
+END_DATE_RAW=${2:-$(date -d "today" +"%Y-%m-%d 00:00")}
+
+# Convert raw date strings to timestamps
+if [[ "$START_DATE_RAW" == *" "* ]]; then
+  START_DATE=$(date -d "$START_DATE_RAW" +"%s")
+else
+  START_DATE=$(date -d "${START_DATE_RAW} 00:00:00" +"%s")
+fi
+
+if [[ "$END_DATE_RAW" == *" "* ]]; then
+  END_DATE=$(date -d "$END_DATE_RAW" +"%s")
+else
+  END_DATE=$(date -d "${END_DATE_RAW} 00:00:00" +"%s")
+fi
+
+echo "Start timestamp: $START_TIMESTAMP"
+echo "End timestamp: $END_TIMESTAMP"
+cd ../
+source .venv/bin/activate
+rm -rf user_data/models/*
+rm -rf ./freqtrade/user_data/data/backtest_results/*
+rm -fr ./user_data/dryrun_results/*
+
+#docker-compose -f docker-compose_backtest.yml run --rm freqtrade >output.log 2>&1
+echo "docker-compose run --rm freqtrade hyperopt \
+  --logfile /freqtrade/user_data/logs/freqtrade.log \
+  --freqaimodel LightGBMRegressorMultiTarget \
+  --strategy $STRATEGY_NAME \
+  --config /freqtrade/config_examples/$CONFIG_FILE \
+  --config /freqtrade/templates/${PARAMS_NAME}.json \
+  --strategy-path /freqtrade/templates \
+  --timerange ${START_DATE}-${END_DATE} \
+  -e 100 \
+  --hyperopt-loss ShortTradeDurHyperOptLoss \
+  --spaces buy sell trailing \
+  --fee 0.0016"
+docker-compose run --rm freqtrade hyperopt \
+  --logfile /freqtrade/user_data/logs/freqtrade.log \
+  --freqaimodel LightGBMRegressorMultiTarget \
+  --strategy $STRATEGY_NAME \
+  --config /freqtrade/config_examples/$CONFIG_FILE \
+  --config /freqtrade/templates/${PARAMS_NAME}.json \
+  --strategy-path /freqtrade/templates \
+  --timerange ${START_DATE}-${END_DATE} \
+  -e 100 \
+  --hyperopt-loss SharpeHyperOptLoss \
+  --spaces buy sell trailing \
+  --fee 0.0016
+
+#>output.log 2>&1
+#sed -i 's/\x1B\[[0-9;]*m//g' output.log
+
+#python3 tools/filter.py
+
+# rm ./result/*.json -fr
+# rm ./result/*.py -fr
+# mv ./user_data/backtest_results/* ./result/
+#
+# cd ./result
+# # 查找当前目录下的所有 zip 文件
+# zip_files=(*.zip)
+#
+# # 检查是否只有一个 zip 文件
+# if [ ${#zip_files[@]} -eq 1 ]; then
+#   # 解压缩该 zip 文件到当前目录
+#   unzip "${zip_files[0]}"
+#   rm *.zip
+#   rm *.feather
+# else
+#   echo "当前目录下没有 zip 文件或者有多个 zip 文件，无法操作。"
+# fi
+#
+# cd -
+# sed -i 's/\x1B\[[0-9;]*m//g' output.log
+# #python3 ../filter.py
+# cp output.log result/ -f
+# cd tools/
+# python tradestocsv.py
+# python analytic.py >../result/analytic.log
+# cd ../

tools/live.sh

@@ -0,0 +1,95 @@
+#!/bin/bash
+
+set -e # 出错立即停止
+# 检查 .env 文件
+if [ ! -f ".env" ]; then
+  echo "⚠️ 本地缺少 .env 文件，请创建并配置。示例内容如下："
+  echo ""
+  echo "STRATEGY_NAME=TheForceV7"
+  echo "CONFIG_FILE=basic.json"
+  echo "TEST_BRANCH=theforce-noai-test"
+  echo "DRYRUN_BRANCH=theforce-noai-dryrun"
+  echo ""
+  exit 1
+fi
+# 加载 .env 文件中的变量（位于项目根目录）
+if [ -f ".env" ]; then
+  export $(grep -v '^#' .env | xargs)
+else
+  echo "⚠️ 缺少 .env 文件，请创建并配置 TEST_BRANCH 和 DRYRUN_BRANCH"
+  exit 1
+fi
+
+# 获取当前分支名
+current_branch=$(git rev-parse --abbrev-ref HEAD)
+
+# 检查当前分支是否匹配 TEST_BRANCH 或 DRYRUN_BRANCH
+if [[ "$current_branch" != "$TEST_BRANCH" && "$current_branch" != "$DRYRUN_BRANCH" ]]; then
+  echo "⚠️ 错误：当前分支 '$current_branch' 不符合环境变量配置要求。"
+  echo "请确保当前分支与 .env 文件中的 TEST_BRANCH 或 DRYRUN_BRANCH 配置一致。"
+  echo "TEST_BRANCH=$TEST_BRANCH"
+  echo "DRYRUN_BRANCH=$DRYRUN_BRANCH"
+  exit 1
+else
+  echo "✅ 当前分支 '$current_branch' 符合环境变量配置要求。"
+fi
+
+# 如果当前分支名称包含 "dryrun"，执行 git reset --hard
+if [[ "$CURRENT_BRANCH" == *"dryrun"* ]]; then
+  echo "当前分支为 '$CURRENT_BRANCH'，正在执行 git reset --hard 以确保代码干净..."
+  git reset --hard
+  if [ $? -ne 0 ]; then
+    echo "⚠️ 执行 git reset --hard 失败，请检查 Git 状态。"
+    exit 1
+  fi
+  echo "Git 工作区已清理。"
+fi
+
+# 加载 .env 文件中的变量
+export $(grep -v '^#' .env | xargs)
+# 设置默认值
+STRATEGY_NAME=${STRATEGY_NAME:-TheForceV7}
+CONFIG_FILE=${CONFIG_FILE:-basic.json}
+# 设置 PARAMS_NAME 为 STRATEGY_NAME 的小写形式
+PARAMS_NAME=$(echo "$STRATEGY_NAME" | tr '[:upper:]' '[:lower:]')
+
+echo "Using strategy: $STRATEGY_NAME"
+echo "Using config: $CONFIG_FILE"
+echo "Using testBranch: $TEST_BRANCH"
+# Parse command line arguments
+START_DATE=${1:-$(date -d "2 days ago" +"%Y%m%d")}
+END_DATE=${2:-$(date -d "tomorrow" +"%Y%m%d")}
+
+cd ../
+source .venv/bin/activate
+rm -rf user_data/models/*
+rm -rf ./freqtrade/user_data/data/backtest_results/*
+rm -fr ./user_data/dryrun_results/*
+rm result/* || true
+
+hyperopt_config="${STRATEGY_NAME%.py}.json"
+docker rm $(docker ps -aq) -f || true
+
+# 获取当前 Git Commit 的前 8 位
+GIT_COMMIT_SHORT=$(git rev-parse HEAD | cut -c 1-8)
+echo "docker-compose run -d --rm --name freqtrade-dryrun-${GIT_COMMIT_SHORT} -p 8080:8080 freqtrade trade \
+  --logfile /freqtrade/user_data/logs/freqtrade.log \
+  --db-url sqlite:////freqtrade/user_data/tradesv3.sqlite \
+  --freqaimodel LightGBMRegressorMultiTarget \
+  --fee 0.0008
+  --config /freqtrade/config_examples/$CONFIG_FILE \
+  --config /freqtrade/templates/${PARAMS_NAME}.json \
+  --config /freqtrade/config_examples/live.json \
+  --strategy $STRATEGY_NAME \
+  --strategy-path /freqtrade/templates"
+
+docker-compose run -d --rm --name freqtrade-dryrun-${GIT_COMMIT_SHORT} -p 8080:8080 freqtrade trade \
+  --logfile /freqtrade/user_data/logs/freqtrade.log \
+  --db-url sqlite:////freqtrade/user_data/tradesv3.sqlite \
+  --freqaimodel LightGBMRegressorMultiTarget \
+  --fee 0.0008 \
+  --config /freqtrade/config_examples/$CONFIG_FILE \
+  --config /freqtrade/templates/${PARAMS_NAME}.json \
+  --config /freqtrade/config_examples/live.json \
+  --strategy $STRATEGY_NAME \
+  --strategy-path /freqtrade/templates

tools/mergeme.sh

@@ -0,0 +1,42 @@
+#!/bin/bash
+
+set -e # 出错立即停止
+
+# 获取参数
+source_branch=$1
+target_branch=$2
+
+# 确保传入了两个参数
+if [ -z "$source_branch" ] || [ -z "$target_branch" ]; then
+  echo "❌ 使用方法: $0 <源分支> <目标分支>"
+  echo "示例: $0 theforce-noai-test base"
+  exit 1
+fi
+
+# 定位到 Git 仓库根目录
+cd "$(git rev-parse --show-toplevel)"
+
+# 获取当前分支名
+current_branch=$(git rev-parse --abbrev-ref HEAD)
+
+# 如果当前不在 source_branch 上，则切换过去
+if [ "$current_branch" != "$source_branch" ]; then
+  echo "🔄 当前在分支 '$current_branch'，正在切换到 '$source_branch'"
+  git checkout "$source_branch"
+fi
+
+echo "📥 正在拉取 '$source_branch' 的最新代码..."
+git pull origin "$source_branch"
+
+echo "📥 正在拉取目标分支 '$target_branch' 的最新代码..."
+git fetch origin "$target_branch"
+
+echo "🔄 正在合并 '$target_branch' 到 '$source_branch'"
+git merge "origin/$target_branch" -m "Merge branch '$target_branch' into $source_branch"
+
+# 如果有冲突，需要手动解决，脚本会在此处中断
+
+echo "📤 正在推送合并后的代码到远程 '$source_branch'"
+git push origin "$source_branch"
+
+echo "✅ 合并完成并已推送到远程分支 '$source_branch'"

tools/mergetemp.sh

@@ -0,0 +1,11 @@
+#!/bin/bash
+
+source_branch=$0
+dest_branch=$1
+path=$3
+cd $path
+git reset --hard
+git checkout $dest_branch
+git pull
+git checkout $source_branch
+git merge $dest_branch

tools/plot.sh

@@ -0,0 +1,11 @@
+#!/bin/bash
+
+docker-compose run --rm freqtrade plot-dataframe \
+  --logfile /freqtrade/user_data/logs/freqtrade.log \
+  --freqaimodel LightGBMRegressorMultiTarget \
+  --config /freqtrade/config_examples/freqaiprimer.json \
+  --strategy-path /freqtrade/templates \
+  --strategy FreqaiPrimer \
+  --timerange 20250410-20250411
+
+

tools/tradestocsv.py

@@ -0,0 +1,50 @@
+import json
+import pandas as pd
+import os
+from pathlib import Path
+
+# 定义结果目录
+result_dir = Path('../result')
+
+# 确保结果目录存在
+if not result_dir.exists():
+    raise FileNotFoundError(f"Directory {result_dir} does not exist")
+
+# 寻找文件名包含 'backtest-result-' 的 JSON 文件
+json_files = [f for f in result_dir.glob('*.json') if 'backtest-result-' in f.name]
+
+if not json_files:
+    raise FileNotFoundError("No JSON files with 'backtest-result-' in name found in ../result")
+
+# 找到文件大小最大的 JSON 文件
+largest_file = max(json_files, key=lambda x: x.stat().st_size)
+
+# 读取最大的 JSON 文件
+with open(largest_file) as f:
+    data = json.load(f)
+
+# 从环境变量中获取策略名称
+strategy_name = os.environ.get('STRATEGY_NAME', 'TheForceV7')  # 如果未设置环境变量，则使用默认值 'TheForceV7'
+
+# 提取交易记录
+trades = data['strategy'][strategy_name]['trades']
+trades_df = pd.DataFrame(trades)
+
+# 选择关键字段
+trades_df = trades_df[[
+    'pair', 'open_date', 'close_date', 'open_rate', 'close_rate', 'amount',
+    'profit_ratio', 'profit_abs', 'exit_reason', 'fee_open', 'fee_close',
+    'trade_duration', 'min_rate', 'max_rate'
+]]
+
+# 转换为本地时间（可选）
+trades_df['open_date'] = pd.to_datetime(trades_df['open_date'])
+trades_df['close_date'] = pd.to_datetime(trades_df['close_date'])
+
+# 定义输出 CSV 文件路径（在 ../result 目录下）
+output_csv = result_dir / 'backtest_trades.csv'
+
+# 保存为 CSV
+trades_df.to_csv(output_csv, index=False)
+
+print(f"Successfully converted {largest_file} to {output_csv}")

tools/view_feather.py

@@ -0,0 +1,65 @@
+import argparse
+import pandas as pd
+
+def analyze_candlestick_data(file_path):
+    # 读取feather文件
+    df = pd.read_feather(file_path)
+
+    # 查看数据集行数和列数
+    rows, columns = df.shape
+
+    if rows < 500:
+        # 短表数据（行数少于500）查看全量数据信息
+        print('数据全部内容信息：')
+        print(df.to_csv(sep='\t', na_rep='nan'))
+    else:
+        # 长表数据查看数据前几行信息
+        print('数据前几行内容信息：')
+        print(df.head().to_csv(sep='\t', na_rep='nan'))
+
+    # 查看数据的基本信息
+    print('数据基本信息：')
+    df.info()
+
+    # 查看数据集行数和列数
+    rows, columns = df.shape
+
+    if columns < 10 and rows < 500:
+        # 短表窄数据（列少于10且行数少于500）查看全量统计信息
+        print('数据全部内容描述性统计信息：')
+        print(df.describe(include='all', percentiles=[.25, .5, .75]).to_csv(sep='\t', na_rep='nan'))
+    else:
+        # 长表数据查看数据前几行统计信息
+        print('数据前几行描述性统计信息：')
+        print(df.head().describe(include='all', percentiles=[.25, .5, .75]).to_csv(sep='\t', na_rep='nan'))
+
+    # 计算时间跨度
+    min_date = df['date'].min()
+    max_date = df['date'].max()
+    time_span = max_date - min_date
+
+    # 检查时间序列完整性
+    df = df.sort_values('date')  # 确保数据按时间排序
+    df['time_diff'] = df['date'].diff().dt.total_seconds()  # 计算相邻时间点的差值(秒)
+    expected_freq = df['time_diff'].mode()[0]  # 使用最常见的间隔作为预期频率
+    missing_intervals = df[df['time_diff'] > expected_freq]  # 找出间隔大于预期的位置
+
+    print(f"\n数据时间跨度：{time_span}")
+    print(f"开始时间：{min_date}")
+    print(f"结束时间：{max_date}")
+
+    if missing_intervals.empty:
+        print("数据完整性：完整，未发现缺失的蜡烛图数据")
+    else:
+        print(f"数据完整性：不完整，发现 {len(missing_intervals)} 处可能的缺失")
+        print("缺失位置示例：")
+        for _, row in missing_intervals.head(5).iterrows():  # 显示前5个缺失示例
+            gap_duration = pd.Timedelta(seconds=row['time_diff'])
+            print(f"  - 在 {row['date']} 之前缺失了 {gap_duration}")
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser(description='分析Freqtrade蜡烛图Feather文件')
+    parser.add_argument('--path', required=True, help='Feather文件路径')
+    args = parser.parse_args()
+    
+    analyze_candlestick_data(args.path)