phyer/myTestFreqAI
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

昨天 经过grok优化的代码跑的很流畅, 没在backtesting里充分验证 就 去 gry-run了, 今天下午backtesting结果大幅跑输,一个月后本金还剩77%, 回退到上一个版本 逐行验证, 动态进退的AI逻辑被改动了, 不知道是grok改的还是qwen-max改的, 然后我改回来了, 现在好了, 以后代码验证机制,流程一定要整明白!

Browse Source
This commit is contained in:
zhangkun9038@dingtalk.com 2025-04-25 19:40:49 +08:00
parent 05122af764
commit ee2926e6a4
4 changed files with 300 additions and 16 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
config_examples/config_freqai.okx.json
View File

@ -70,7 +70,7 @@
},
"freqaimodel": "CatboostClassifier",
"purge_old_models": 2,
"identifier": "test131",
"identifier": "test158",
"train_period_days": 30,
"backtest_period_days": 10,
"live_retrain_hours": 0,

4
docker-compose.yml
View File

@ -40,7 +40,7 @@ services:
--config /freqtrade/config_examples/config_freqai.okx.json
--strategy FreqaiExampleStrategy
--strategy-path /freqtrade/templates
--fee 0.001
--fee 0.0008
# command: >
# backtesting
# --logfile /freqtrade/user_data/logs/freqtrade.log
@ -71,7 +71,7 @@ services:
# --strategy-path /freqtrade/templates
# --strategy FreqaiExampleStrategy
# --timerange 20250301-20250420
# --fee 0.001
# --fee 0.0008
# command: >
# hyperopt

262
freqtrade/templates/FreqaiExampleStrategy.diff Normal file
View File

@ -0,0 +1,262 @@
diff --git a/freqtrade/templates/FreqaiExampleStrategy.py b/freqtrade/templates/FreqaiExampleStrategy.py
index 343c073..1d7ed33 100644
--- a/freqtrade/templates/FreqaiExampleStrategy.py
+++ b/freqtrade/templates/FreqaiExampleStrategy.py
@@ -11,9 +11,9 @@ logger = logging.getLogger(__name__)
class FreqaiExampleStrategy(IStrategy):
minimal_roi = {
- "0": 0.076,
- "7": 0.034,
- "13": 0.007,
+ "0": 0.02,
+ "7": 0.01,
+ "13": 0.005,
"60": 0
}
@@ -24,29 +24,25 @@ class FreqaiExampleStrategy(IStrategy):
startup_candle_count: int = 40
can_short = False
- # Hyperopt 参数
buy_rsi = IntParameter(low=10, high=50, default=27, space="buy", optimize=True, load=True)
sell_rsi = IntParameter(low=50, high=90, default=59, space="sell", optimize=True, load=True)
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=-0.25, high=-0.05, default=-0.1, space="stoploss", optimize=True, load=True)
- trailing_stop_positive_offset = DecimalParameter(low=0.01, high=0.5, default=0.02, space="trailing", optimize=True, load=True)
+ trailing_stop_positive_offset = DecimalParameter(low=0.005, high=0.5, default=0.01, space="trailing", optimize=True, load=True)
- protections = [
- {"method": "StoplossGuard", "stop_duration": 60, "lookback_period": 120},
- {"method": "MaxDrawdown", "lookback_period": 120, "max_allowed_drawdown": 0.05}
- ]
+ protections = []
freqai_info = {
"model": "LightGBMRegressor",
"feature_parameters": {
"include_timeframes": ["5m"],
- "include_corr_pairlist": ["SOL/USDT", "BTC/USDT"],
+ "include_corr_pairlist": ["SOL/USDT"],
"label_period_candles": 12,
"include_shifted_candles": 0,
- "include_periods": [10, 20],
- "DI_threshold": 3.0
+ "include_periods": [20],
+ "DI_threshold": 5.0
},
"data_split_parameters": {
"test_size": 0.2,
@@ -62,14 +58,20 @@ class FreqaiExampleStrategy(IStrategy):
}
plot_config = {
- "main_plot": {},
+ "main_plot": {
+ "close": {"color": "blue"},
+ "bb_lowerband": {"color": "purple"}
+ },
"subplots": {
"&-buy_rsi": {"&-buy_rsi": {"color": "green"}},
"&-sell_rsi": {"&-sell_rsi": {"color": "red"}},
- "&-stoploss": {"&-stoploss": {"color": "purple"}},
- "&-roi_0": {"&-roi_0": {"color": "orange"}},
+ "rsi": {"rsi": {"color": "black"}},
"do_predict": {"do_predict": {"color": "brown"}},
- },
+ "trade_signals": {
+ "enter_long": {"color": "green", "type": "scatter"},
+ "exit_long": {"color": "red", "type": "scatter"}
+ }
+ }
}
def feature_engineering_expand_all(self, dataframe: DataFrame, period: int, metadata: dict, **kwargs) -> DataFrame:
@@ -130,12 +132,10 @@ class FreqaiExampleStrategy(IStrategy):
logger.info(f"DataFrame rows: {len(dataframe)}")
logger.info(f"Columns before freqai.start: {list(dataframe.columns)}")
- # 验证输入数据
if "close" not in dataframe.columns or dataframe["close"].isna().all():
logger.error(f"DataFrame missing 'close' column or all NaN for pair: {metadata['pair']}")
raise ValueError("DataFrame missing valid 'close' column")
- # 生成 RSI
if len(dataframe) < 14:
logger.warning(f"DataFrame too short ({len(dataframe)} rows), cannot compute rsi")
dataframe["rsi"] = 50
@@ -143,7 +143,6 @@ class FreqaiExampleStrategy(IStrategy):
dataframe["rsi"] = ta.RSI(dataframe, timeperiod=14)
logger.info(f"rsi stats: {dataframe['rsi'].describe().to_string()}")
- # 生成 %-volatility
if len(dataframe) < 20 or dataframe["close"].isna().any():
logger.warning(f"DataFrame too short ({len(dataframe)} rows) or contains NaN in close, cannot compute %-volatility")
dataframe["%-volatility"] = 0
@@ -154,7 +153,6 @@ class FreqaiExampleStrategy(IStrategy):
dataframe["%-volatility"] = (dataframe["%-volatility"] - dataframe["%-volatility"].mean()) / dataframe["%-volatility"].std()
logger.info(f"%-volatility stats: {dataframe['%-volatility'].describe().to_string()}")
- # 生成 TEMA
if len(dataframe) < 9:
logger.warning(f"DataFrame too short ({len(dataframe)} rows), cannot compute tema")
dataframe["tema"] = dataframe["close"]
@@ -165,7 +163,6 @@ class FreqaiExampleStrategy(IStrategy):
dataframe["tema"] = dataframe["tema"].fillna(dataframe["close"])
logger.info(f"tema stats: {dataframe['tema'].describe().to_string()}")
- # 生成 Bollinger Bands
if len(dataframe) < 20:
logger.warning(f"DataFrame too short ({len(dataframe)} rows), cannot compute bb_lowerband")
dataframe["bb_lowerband"] = dataframe["close"]
@@ -177,21 +174,6 @@ class FreqaiExampleStrategy(IStrategy):
dataframe["bb_lowerband"] = dataframe["bb_lowerband"].fillna(dataframe["close"])
logger.info(f"bb_lowerband stats: {dataframe['bb_lowerband'].describe().to_string()}")
- # 生成 up_or_down
- label_period = self.freqai_info["feature_parameters"]["label_period_candles"]
- if len(dataframe) < label_period + 1:
- logger.warning(f"DataFrame too short ({len(dataframe)} rows), cannot compute up_or_down")
- dataframe["up_or_down"] = 0
- else:
- dataframe["up_or_down"] = np.where(
- dataframe["close"].shift(-label_period) > dataframe["close"], 1, 0
- )
- if dataframe["up_or_down"].isna().any():
- logger.warning("up_or_down contains NaN, filling with 0")
- dataframe["up_or_down"] = dataframe["up_or_down"].fillna(0)
- logger.info(f"up_or_down stats: {dataframe['up_or_down'].describe().to_string()}")
-
- # 生成其他特征
if "date" in dataframe.columns:
dataframe["%-day_of_week"] = dataframe["date"].dt.dayofweek
dataframe["%-hour_of_day"] = dataframe["date"].dt.hour
@@ -200,7 +182,6 @@ class FreqaiExampleStrategy(IStrategy):
dataframe["%-day_of_week"] = 0
dataframe["%-hour_of_day"] = 0
- # 调用 FreqAI
try:
dataframe = self.freqai.start(dataframe, metadata, self)
logger.info(f"Columns after freqai.start: {list(dataframe.columns)}")
@@ -210,26 +191,23 @@ class FreqaiExampleStrategy(IStrategy):
dataframe["sell_rsi_pred"] = 80
dataframe["do_predict"] = 1
- # 检查预测列
for col in ["buy_rsi_pred", "sell_rsi_pred"]:
if col not in dataframe.columns:
logger.error(f"Error: {col} column not generated for pair: {metadata['pair']}")
dataframe[col] = 50 if col == "buy_rsi_pred" else 80
logger.info(f"{col} stats: {dataframe[col].describe().to_string()}")
- # 调试特征分布
- if "%-bb_width-period_10_SOL/USDT_5m" in dataframe.columns:
- if dataframe["%-bb_width-period_10_SOL/USDT_5m"].std() > 0:
- dataframe["%-bb_width-period_10_SOL/USDT_5m"] = (
- dataframe["%-bb_width-period_10_SOL/USDT_5m"] - dataframe["%-bb_width-period_10_SOL/USDT_5m"].mean()
- ) / dataframe["%-bb_width-period_10_SOL/USDT_5m"].std()
- logger.info(f"%-bb_width-period_10 stats: {dataframe['%-bb_width-period_10_SOL/USDT_5m'].describe().to_string()}")
+ if "%-bb_width-period_20_SOL/USDT_5m" in dataframe.columns:
+ if dataframe["%-bb_width-period_20_SOL/USDT_5m"].std() > 0:
+ dataframe["%-bb_width-period_20_SOL/USDT_5m"] = (
+ dataframe["%-bb_width-period_20_SOL/USDT_5m"] - dataframe["%-bb_width-period_20_SOL/USDT_5m"].mean()
+ ) / dataframe["%-bb_width-period_20_SOL/USDT_5m"].std()
+ logger.info(f"%-bb_width-period_20 stats: {dataframe['%-bb_width-period_20_SOL/USDT_5m'].describe().to_string()}")
- # 动态生成期望的特征列
def get_expected_columns(freqai_config: dict) -> list:
indicators = ["rsi", "bb_width", "pct-change"]
- periods = freqai_config.get("feature_parameters", {}).get("include_periods", [10, 20])
- pairs = freqai_config.get("include_corr_pairlist", ["SOL/USDT", "BTC/USDT"])
+ periods = freqai_config.get("feature_parameters", {}).get("include_periods", [20])
+ pairs = freqai_config.get("include_corr_pairlist", ["SOL/USDT"])
timeframes = freqai_config.get("include_timeframes", ["5m"])
shifts = [0]
expected_columns = ["%-volatility", "%-day_of_week", "%-hour_of_day"]
@@ -248,50 +226,47 @@ class FreqaiExampleStrategy(IStrategy):
expected_columns = get_expected_columns(self.freqai_info)
logger.info(f"Expected feature columns ({len(expected_columns)}): {expected_columns[:10]}...")
- # 比较特征集
actual_columns = list(dataframe.columns)
missing_columns = [col for col in expected_columns if col not in actual_columns]
extra_columns = [col for col in actual_columns if col not in expected_columns and col.startswith("%-")]
logger.info(f"Missing columns ({len(missing_columns)}): {missing_columns}")
logger.info(f"Extra columns ({len(extra_columns)}): {extra_columns}")
- # 调试 DI 丢弃预测
if "DI_values" in dataframe.columns:
logger.info(f"DI_values stats: {dataframe['DI_values'].describe().to_string()}")
logger.info(f"DI discarded predictions: {len(dataframe[dataframe['do_predict'] == 0])}")
- # 清理数据
dataframe = dataframe.replace([np.inf, -np.inf], 0).ffill().fillna(0)
logger.info(f"Final columns in populate_indicators: {list(dataframe.columns)}")
return dataframe
def populate_entry_trend(self, df: DataFrame, metadata: dict) -> DataFrame:
enter_long_conditions = [
- qtpylib.crossed_above(df["rsi"], df["buy_rsi_pred"] + (5 if metadata["pair"] == "BTC/USDT" else 0)),
+ qtpylib.crossed_above(df["rsi"], df["buy_rsi_pred"]),
df["tema"] > df["tema"].shift(1),
df["volume"] > 0,
- df["do_predict"] == 1,
- df["up_or_down"] == 1
+ df["do_predict"] == 1
]
- if enter_long_conditions:
- df.loc[
- reduce(lambda x, y: x & y, enter_long_conditions),
- ["enter_long", "enter_tag"]
- ] = (1, "long")
+ df["entry_signal"] = reduce(lambda x, y: x & y, enter_long_conditions)
+ df["entry_signal"] = df["entry_signal"].rolling(window=2, min_periods=1).max().astype(bool)
+ df.loc[
+ df["entry_signal"],
+ ["enter_long", "enter_tag"]
+ ] = (1, "long")
+ if df["entry_signal"].iloc[-1]:
+ logger.info(f"Entry signal triggered for {metadata['pair']}: rsi={df['rsi'].iloc[-1]}, buy_rsi_pred={df['buy_rsi_pred'].iloc[-1]}, do_predict={df['do_predict'].iloc[-1]}")
return df
def populate_exit_trend(self, df: DataFrame, metadata: dict) -> DataFrame:
exit_long_conditions = [
- (qtpylib.crossed_above(df["rsi"], df["sell_rsi_pred"])) |
+ (qtpylib.crossed_above(df["rsi"], df["sell_rsi_pred"] - 5)) |
(df["close"] < df["close"].shift(1) * 0.98) |
(df["close"] < df["bb_lowerband"]),
df["volume"] > 0,
- df["do_predict"] == 1,
- df["up_or_down"] == 0
+ df["do_predict"] == 1
]
- time_exit = (df["date"] >= df["date"].shift(1) + pd.Timedelta(days=1))
df.loc[
- (reduce(lambda x, y: x & y, exit_long_conditions)) | time_exit,
+ reduce(lambda x, y: x & y, exit_long_conditions),
"exit_long"
] = 1
return df
@@ -300,9 +275,16 @@ 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:
+ logger.info(f"Confirming trade entry for {pair}, order_type: {order_type}, rate: {rate}, current_time: {current_time}")
df, _ = self.dp.get_analyzed_dataframe(pair, self.timeframe)
last_candle = df.iloc[-1].squeeze()
if side == "long":
- if rate > (last_candle["close"] * (1 + 0.001)):
- return False
+ if order_type == "market":
+ logger.info(f"Order confirmed for {pair}, rate: {rate} (market order)")
+ return True
+ if rate <= (last_candle["close"] * (1 + 0.01)):
+ logger.info(f"Order confirmed for {pair}, rate: {rate}")
+ return True
+ logger.info(f"Order rejected: rate {rate} exceeds threshold {last_candle['close'] * 1.01}")
+ return False
return True

48
freqtrade/templates/FreqaiExampleStrategy.py
View File

@ -173,6 +173,19 @@ class FreqaiExampleStrategy(IStrategy):
logger.warning("bb_lowerband contains NaN, filling with close")
dataframe["bb_lowerband"] = dataframe["bb_lowerband"].fillna(dataframe["close"])
logger.info(f"bb_lowerband stats: {dataframe['bb_lowerband'].describe().to_string()}")
# 生成 up_or_down
label_period = self.freqai_info["feature_parameters"]["label_period_candles"]
if len(dataframe) < label_period + 1:
logger.warning(f"DataFrame too short ({len(dataframe)} rows), cannot compute up_or_down")
dataframe["up_or_down"] = 0
else:
dataframe["up_or_down"] = np.where(
dataframe["close"].shift(-label_period) > dataframe["close"], 1, 0
)
if dataframe["up_or_down"].isna().any():
logger.warning("up_or_down contains NaN, filling with 0")
dataframe["up_or_down"] = dataframe["up_or_down"].fillna(0)
logger.info(f"up_or_down stats: {dataframe['up_or_down'].describe().to_string()}")
if "date" in dataframe.columns:
dataframe["%-day_of_week"] = dataframe["date"].dt.dayofweek
@ -197,17 +210,18 @@ class FreqaiExampleStrategy(IStrategy):
dataframe[col] = 50 if col == "buy_rsi_pred" else 80
logger.info(f"{col} stats: {dataframe[col].describe().to_string()}")
if "%-bb_width-period_20_SOL/USDT_5m" in dataframe.columns:
if dataframe["%-bb_width-period_20_SOL/USDT_5m"].std() > 0:
dataframe["%-bb_width-period_20_SOL/USDT_5m"] = (
dataframe["%-bb_width-period_20_SOL/USDT_5m"] - dataframe["%-bb_width-period_20_SOL/USDT_5m"].mean()
) / dataframe["%-bb_width-period_20_SOL/USDT_5m"].std()
logger.info(f"%-bb_width-period_20 stats: {dataframe['%-bb_width-period_20_SOL/USDT_5m'].describe().to_string()}")
# 调试特征分布
if "%-bb_width-period_10_SOL/USDT_5m" in dataframe.columns:
if dataframe["%-bb_width-period_10_SOL/USDT_5m"].std() > 0:
dataframe["%-bb_width-period_10_SOL/USDT_5m"] = (
dataframe["%-bb_width-period_10_SOL/USDT_5m"] - dataframe["%-bb_width-period_10_SOL/USDT_5m"].mean()
) / dataframe["%-bb_width-period_10_SOL/USDT_5m"].std()
logger.info(f"%-bb_width-period_10 stats: {dataframe['%-bb_width-period_10_SOL/USDT_5m'].describe().to_string()}")
def get_expected_columns(freqai_config: dict) -> list:
indicators = ["rsi", "bb_width", "pct-change"]
periods = freqai_config.get("feature_parameters", {}).get("include_periods", [20])
pairs = freqai_config.get("include_corr_pairlist", ["SOL/USDT"])
periods = freqai_config.get("feature_parameters", {}).get("include_periods", [10, 20])
pairs = freqai_config.get("include_corr_pairlist", ["SOL/USDT", "BTC/USDT"])
timeframes = freqai_config.get("include_timeframes", ["5m"])
shifts = [0]
expected_columns = ["%-volatility", "%-day_of_week", "%-hour_of_day"]
@ -242,11 +256,17 @@ class FreqaiExampleStrategy(IStrategy):
def populate_entry_trend(self, df: DataFrame, metadata: dict) -> DataFrame:
enter_long_conditions = [
qtpylib.crossed_above(df["rsi"], df["buy_rsi_pred"]),
qtpylib.crossed_above(df["rsi"], df["buy_rsi_pred"] + (5 if metadata["pair"] == "BTC/USDT" else 0)),
df["tema"] > df["tema"].shift(1),
df["volume"] > 0,
df["do_predict"] == 1
df["do_predict"] == 1,
df["up_or_down"] == 1
]
if enter_long_conditions:
df.loc[
reduce(lambda x, y: x & y, enter_long_conditions),
["enter_long", "enter_tag"]
] = (1, "long")
df["entry_signal"] = reduce(lambda x, y: x & y, enter_long_conditions)
df["entry_signal"] = df["entry_signal"].rolling(window=2, min_periods=1).max().astype(bool)
df.loc[
@ -259,14 +279,16 @@ class FreqaiExampleStrategy(IStrategy):
def populate_exit_trend(self, df: DataFrame, metadata: dict) -> DataFrame:
exit_long_conditions = [
(qtpylib.crossed_above(df["rsi"], df["sell_rsi_pred"] - 5)) |
(qtpylib.crossed_above(df["rsi"], df["sell_rsi_pred"])) |
(df["close"] < df["close"].shift(1) * 0.98) |
(df["close"] < df["bb_lowerband"]),
df["volume"] > 0,
df["do_predict"] == 1
df["do_predict"] == 1,
df["up_or_down"] == 0
]
time_exit = (df["date"] >= df["date"].shift(1) + pd.Timedelta(days=1))
df.loc[
reduce(lambda x, y: x & y, exit_long_conditions),
(reduce(lambda x, y: x & y, exit_long_conditions)) | time_exit,
"exit_long"
] = 1
return df