"""技术指标（单一事实源）。

MVP 用纯 pandas/numpy 实现，避免 Windows 上 TA-Lib C 库的安装痛点。
算法正确（MACD = 快慢 EMA 之差），接口稳定；阶段1 在 Linux/Docker 上可换 TA-Lib，
只需保持函数签名（输入 close Series，输出指标），上层无感。
"""
from __future__ import annotations

import numpy as np
import pandas as pd


def ema(series: pd.Series, span: int) -> pd.Series:
    """指数移动平均（adjust=False，与 TA-Lib 默认一致）。"""
    return series.ewm(span=span, adjust=False).mean()


def macd(close: pd.Series, fast: int = 12, slow: int = 26, signal: int = 9) -> pd.DataFrame:
    """MACD：返回 DataFrame[DIF, DEA, HIST]。"""
    dif = ema(close, fast) - ema(close, slow)
    dea = ema(dif, signal)
    hist = (dif - dea) * 2  # A股惯例 MACD 柱 = 2*(DIF-DEA)
    return pd.DataFrame({"macd": dif, "signal": dea, "hist": hist})


def rsi(close: pd.Series, period: int = 14) -> pd.Series:
    """RSI（Wilder 平滑）。"""
    delta = close.diff()
    gain = delta.clip(lower=0.0)
    loss = -delta.clip(upper=0.0)
    avg_gain = gain.ewm(alpha=1 / period, adjust=False).mean()
    avg_loss = loss.ewm(alpha=1 / period, adjust=False).mean()
    rs = avg_gain / avg_loss.replace(0, np.nan)
    return 100 - (100 / (1 + rs))


def kdj(high: pd.Series, low: pd.Series, close: pd.Series,
        n: int = 9, m1: int = 3, m2: int = 3) -> pd.DataFrame:
    """KDJ（A股常用：RSV -> K -> D -> J）。"""
    low_n = low.rolling(n, min_periods=1).min()
    high_n = high.rolling(n, min_periods=1).max()
    rsv = (close - low_n) / (high_n - low_n).replace(0, np.nan) * 100
    k = rsv.ewm(alpha=1 / m1, adjust=False).mean()
    d = k.ewm(alpha=1 / m2, adjust=False).mean()
    j = 3 * k - 2 * d
    return pd.DataFrame({"k": k, "d": d, "j": j})


def bollinger(close: pd.Series, period: int = 20, std: float = 2.0) -> pd.DataFrame:
    ma = close.rolling(period, min_periods=1).mean()
    sd = close.rolling(period, min_periods=1).std(ddof=0)
    return pd.DataFrame({"mid": ma, "upper": ma + std * sd, "lower": ma - std * sd})


def ma(close: pd.Series, period: int) -> pd.Series:
    return close.rolling(period, min_periods=1).mean()