feat(engine): 添加数据服务层与技术指标库

- data/service.py: 数据拉取服务，从 TimescaleDB 读取 K 线/Ticker 等行情数据
- indicators/momentum.py: 动量类指标（RSI/MACD/Stochastic 等）
- indicators/trend.py: 趋势类指标（EMA/SMA/ADX/SuperTrend 等）
- indicators/volatility.py: 波动率指标（Bollinger/ATR/Keltner 等）
- indicators/volume.py: 成交量指标（OBV/VWAP/MFI 等）

engine/indicators/volatility.py

@@ -0,0 +1,127 @@
+"""
+波动率指标 — 布林带、ATR
+
+所有函数返回与输入等长的 list[float]，不足周期位置填 0.0。
+"""
+
+import math
+
+
+def bollinger(
+    data: list[float],
+    period: int = 20,
+    std: float = 2.0,
+):
+    """布林带 (Bollinger Bands)
+
+    使用流式计算方差，O(n) 复杂度。
+
+    Args:
+        data: 价格序列（通常为收盘价）
+        period: 中轨 SMA 周期
+        std: 标准差倍数
+
+    Returns:
+        (upper, mid, lower) 三个等长序列
+    """
+    n = len(data)
+    upper = [0.0] * n
+    mid = [0.0] * n
+    lower = [0.0] * n
+
+    if n < period:
+        return upper, mid, lower
+
+    # 初始窗口的 sum 和 sum_sq
+    window_sum = 0.0
+    window_sum_sq = 0.0
+    for i in range(period):
+        v = data[i]
+        window_sum += v
+        window_sum_sq += v * v
+
+    # 第一个点
+    mean = window_sum / period
+    mid[period - 1] = mean
+    variance = (window_sum_sq / period) - (mean * mean)
+    stdev = math.sqrt(max(variance, 0.0))
+    upper[period - 1] = mean + std * stdev
+    lower[period - 1] = mean - std * stdev
+
+    # 滑动窗口计算后续点
+    for i in range(period, n):
+        old_val = data[i - period]
+        new_val = data[i]
+        window_sum += new_val - old_val
+        window_sum_sq += new_val * new_val - old_val * old_val
+
+        mean = window_sum / period
+        mid[i] = mean
+        variance = (window_sum_sq / period) - (mean * mean)
+        stdev = math.sqrt(max(variance, 0.0))
+        upper[i] = mean + std * stdev
+        lower[i] = mean - std * stdev
+
+    return upper, mid, lower
+
+
+def bollinger_upper(data: list[float], period: int = 20, std: float = 2.0) -> list[float]:
+    """布林带上轨"""
+    upper, _, _ = bollinger(data, period, std)
+    return upper
+
+
+def bollinger_mid(data: list[float], period: int = 20) -> list[float]:
+    """布林带中轨"""
+    from .trend import sma as _sma
+    return _sma(data, period)
+
+
+def bollinger_lower(data: list[float], period: int = 20, std: float = 2.0) -> list[float]:
+    """布林带下轨"""
+    _, _, lower = bollinger(data, period, std)
+    return lower
+
+
+def atr(
+    high: list[float],
+    low: list[float],
+    close: list[float],
+    period: int = 14,
+) -> list[float]:
+    """平均真实波幅 (ATR)
+
+    使用 Wilder 平滑算法。
+
+    Args:
+        high: 最高价序列
+        low: 最低价序列
+        close: 收盘价序列
+        period: 周期
+
+    Returns:
+        与输入等长的 ATR 序列，前 period 位置为 0
+    """
+    n = len(close)
+    result = [0.0] * n
+    if n < period + 1:
+        return result
+
+    # 计算 True Range
+    tr = [0.0] * n
+    tr[0] = high[0] - low[0]
+    for i in range(1, n):
+        tr[i] = max(
+            high[i] - low[i],
+            abs(high[i] - close[i - 1]),
+            abs(low[i] - close[i - 1]),
+        )
+
+    # 初始 ATR 为前 period 个 TR 的均值
+    result[period] = sum(tr[1:period + 1]) / period
+
+    # Wilder 平滑
+    for i in range(period + 1, n):
+        result[i] = (result[i - 1] * (period - 1) + tr[i]) / period
+
+    return result