chore(config): 添加环境变量示例及.gitignore更新

- 新增 .env.example，包含 Tushare API、钉钉机器人和PostgreSQL数据库配置模板
- 更新.gitignore，忽略本地配置文件如 .env.local 和 config_local.py
- 添加对报表文件命名规则的支持，保留示例文件不忽略
- 删除废弃的 chart.py 及相关图表模块代码
- 新增 config/settings.py，实现从环境变量读取配置的统一接口
- 设置数据目录及缓存目录，确保目录存在，提高配置管理规范性

core/factors/momentum.py

@@ -0,0 +1,137 @@
+"""
+动量因子计算模块
+
+支持两种动量因子：
+1. N日涨幅（简单动量）
+2. 斜率×R²趋势得分（改进版）
+"""
+
+import numpy as np
+import pandas as pd
+from sklearn.linear_model import LinearRegression
+
+
+def calculate_momentum(price_series: pd.Series, n: int) -> pd.Series:
+    """
+    计算 N 日涨幅（简单动量）
+
+    Args:
+        price_series: 价格序列
+        n: 动量窗口天数
+
+    Returns:
+        Series: N日涨幅
+    """
+    return price_series / price_series.shift(n + 1) - 1.0
+
+
+def _slope_r2_score(srs: pd.Series, n: int = 25) -> float:
+    """
+    单次计算斜率×R²趋势得分
+
+    Args:
+        srs: 价格窗口序列（长度为 n）
+        n: 窗口长度
+
+    Returns:
+        float: 斜率 × R² × 10000
+    """
+    if srs.shape[0] < n:
+        return np.nan
+
+    x = np.arange(1, n + 1).reshape(-1, 1)
+    y = srs.values / srs.values[0]  # 归一化
+
+    lr = LinearRegression().fit(x, y)
+    slope = lr.coef_[0]
+    r_squared = lr.score(x, y)
+    score = 10000 * slope * r_squared
+
+    return score
+
+
+def calculate_slope_r2(price_series: pd.Series, n: int = 25) -> pd.Series:
+    """
+    计算斜率×R²趋势得分序列
+
+    Args:
+        price_series: 价格序列
+        n: 滚动窗口天数
+
+    Returns:
+        Series: 趋势得分序列
+    """
+    return price_series.rolling(n).apply(
+        lambda x: _slope_r2_score(x, n), raw=False
+    )
+
+
+def calculate_daily_return(price_series: pd.Series) -> pd.Series:
+    """
+    计算日收益率
+
+    Args:
+        price_series: 价格序列
+
+    Returns:
+        Series: 日收益率
+    """
+    return price_series / price_series.shift(1) - 1
+
+
+def compute_factors(
+    etf_data: pd.DataFrame,
+    code_list: list,
+    n: int = 25,
+    factor_type: str = "slope_r2",
+) -> tuple[pd.DataFrame, list]:
+    """
+    计算所有指数的因子和日收益率
+
+    Args:
+        etf_data: DataFrame, 宽表格式的收盘价
+        code_list: 指数代码列表
+        n: 动量/趋势窗口
+        factor_type: 'momentum' 或 'slope_r2'
+
+    Returns:
+        tuple: (result_df, valid_codes)
+    """
+    result = etf_data.copy()
+
+    # 过滤掉缺失值过多的指数
+    total_rows = len(result)
+    valid_codes = []
+    for code in code_list:
+        if code not in result.columns:
+            print(f"  ⚠ 跳过 {code}: 不在数据中")
+            continue
+        null_pct = result[code].isnull().sum() / total_rows
+        if null_pct > 0.2:
+            print(f"  ⚠ 剔除 {code}: 缺失率 {null_pct:.1%} 过高")
+            result = result.drop(columns=[code])
+        else:
+            valid_codes.append(code)
+
+    # 对有效指数计算因子
+    for code in valid_codes:
+        result[f"日收益率_{code}"] = calculate_daily_return(result[code])
+
+        if factor_type == "momentum":
+            result[f"得分_{code}"] = calculate_momentum(result[code], n)
+        elif factor_type == "slope_r2":
+            result[f"得分_{code}"] = calculate_slope_r2(result[code], n)
+        else:
+            raise ValueError(f"不支持的因子类型: {factor_type}")
+
+    # 按得分列做 dropna
+    score_cols = [f"得分_{code}" for code in valid_codes]
+    result = result.dropna(subset=score_cols)
+
+    print("\n因子计算完成:")
+    print(f"  因子类型: {factor_type}")
+    print(f"  窗口天数: {n}")
+    print(f"  有效指数: {len(valid_codes)}/{len(code_list)}")
+    print(f"  有效数据: {len(result)} 行")
+
+    return result, valid_codes

core/factors/technical.py

@@ -0,0 +1,207 @@
+"""
+技术指标计算模块
+
+包含CCI、EMA、MACD等常用技术指标
+"""
+
+import pandas as pd
+import numpy as np
+import talib as ta
+
+
+def calculate_cci(
+    df: pd.DataFrame,
+    period: int = 14,
+    high_col: str = "high",
+    low_col: str = "low",
+    close_col: str = "close",
+) -> pd.Series:
+    """
+    计算CCI指标（商品通道指数）
+
+    Args:
+        df: DataFrame with OHLC data
+        period: CCI周期
+        high_col: 最高价列名
+        low_col: 最低价列名
+        close_col: 收盘价列名
+
+    Returns:
+        Series: CCI值
+    """
+    return ta.CCI(
+        high=df[high_col],
+        low=df[low_col],
+        close=df[close_col],
+        timeperiod=period,
+    )
+
+
+def calculate_ema(
+    price_series: pd.Series,
+    period: int = 20,
+) -> pd.Series:
+    """
+    计算指数移动平均线
+
+    Args:
+        price_series: 价格序列
+        period: EMA周期
+
+    Returns:
+        Series: EMA值
+    """
+    return ta.EMA(price_series, timeperiod=period)
+
+
+def calculate_macd(
+    price_series: pd.Series,
+    fastperiod: int = 12,
+    slowperiod: int = 26,
+    signalperiod: int = 9,
+) -> tuple[pd.Series, pd.Series, pd.Series]:
+    """
+    计算MACD指标
+
+    Args:
+        price_series: 价格序列
+        fastperiod: 快线周期
+        slowperiod: 慢线周期
+        signalperiod: 信号线周期
+
+    Returns:
+        tuple: (macd, signal, hist)
+    """
+    macd, signal, hist = ta.MACD(
+        price_series,
+        fastperiod=fastperiod,
+        slowperiod=slowperiod,
+        signalperiod=signalperiod,
+    )
+    return macd, signal, hist
+
+
+def calculate_td_sequence(close_series: pd.Series) -> pd.Series:
+    """
+    计算TD序列（Tom DeMark Sequential）
+
+    Args:
+        close_series: 收盘价序列
+
+    Returns:
+        Series: TD序列值（正数为上涨计数，负数为下跌计数）
+    """
+    close = close_series.to_list()
+    td = [0, 0, 0, 0]
+    up = 0
+    down = 0
+
+    for i in range(4, len(close)):
+        if close[i] > close[i - 4]:
+            up += 1
+            down = 0
+            td.append(up)
+        else:
+            down -= 1
+            up = 0
+            td.append(down)
+
+    return pd.Series(td, index=close_series.index)
+
+
+def resample_to_weekly(df: pd.DataFrame) -> pd.DataFrame:
+    """
+    将日线数据重采样为周线数据
+
+    Args:
+        df: DataFrame with columns: date, open, high, low, close, volume
+
+    Returns:
+        DataFrame: 周线数据
+    """
+    df = df.copy()
+    if "date" in df.columns:
+        df["date"] = pd.to_datetime(df["date"])
+        df.set_index("date", inplace=True)
+
+    weekly = pd.DataFrame(
+        {
+            "code": df["code"].resample("W").first() if "code" in df.columns else None,
+            "open": df["open"].resample("W").first(),
+            "high": df["high"].resample("W").max(),
+            "low": df["low"].resample("W").min(),
+            "close": df["close"].resample("W").last(),
+            "volume": df["volume"].resample("W").sum(),
+        }
+    )
+
+    return weekly.dropna()
+
+
+class TechnicalScreener:
+    """技术指标筛选器基类"""
+
+    def __init__(self, name: str):
+        self.name = name
+
+    def screen(self, df: pd.DataFrame) -> bool:
+        """
+        判断数据是否符合筛选条件
+
+        Args:
+            df: DataFrame with OHLCV data
+
+        Returns:
+            bool: 是否符合条件
+        """
+        raise NotImplementedError
+
+
+class CCIScreener(TechnicalScreener):
+    """CCI超卖筛选器"""
+
+    def __init__(
+        self,
+        day_period: int = 14,
+        week_period: int = 14,
+        threshold: float = -100,
+        use_weekly: bool = True,
+    ):
+        super().__init__("CCI超卖筛选")
+        self.day_period = day_period
+        self.week_period = week_period
+        self.threshold = threshold
+        self.use_weekly = use_weekly
+
+    def screen(self, df: pd.DataFrame) -> dict:
+        """
+        筛选CCI超卖信号
+
+        Returns:
+            dict: {
+                'triggered': bool,  # 是否触发信号
+                'day_cci': float,   # 日线CCI值
+                'week_cci': float,  # 周线CCI值（如启用）
+            }
+        """
+        # 计算日线CCI
+        day_cci = calculate_cci(df, period=self.day_period).iloc[-1]
+
+        result = {
+            "triggered": day_cci < self.threshold,
+            "day_cci": day_cci,
+            "week_cci": None,
+        }
+
+        # 计算周线CCI（如果启用）
+        if self.use_weekly:
+            weekly_df = resample_to_weekly(df)
+            if len(weekly_df) >= self.week_period:
+                week_cci = calculate_cci(weekly_df, period=self.week_period).iloc[-1]
+                result["week_cci"] = week_cci
+                # 日线或周线任一超卖即触发
+                result["triggered"] = (
+                    day_cci < self.threshold or week_cci < self.threshold
+                )
+
+        return result