python view_feather.py --path ../user_data/data/okx/TRUMP_USDT-5m.feather

tools/download.sh

@@ -60,11 +60,14 @@ if [[ "$@" == *"--timerange"* ]] && [[ "$@" == *"--days"* ]]; then
 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=$(get_csv_param_value "--timeframe" "$@")
@@ -75,18 +78,22 @@ if [[ -z "$pairs" ]]; then
 fi
 
 if [[ -z "$timeframe" ]]; then
-  timeframe="3m,5m,15m,30m,1h,4h,6h,12h,1d"
+  timeframe="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"
+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 $timerange"
+  cmd+=" --timerange='$timerange'"
 elif [[ -n "$days" ]]; then
-  cmd+=" --days $days"
+  cmd+=" --days=$days"
 fi
 
 # Execute the command
-eval $cmd
+eval "$cmd"

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)