import pandas as pd
from sqlalchemy import create_engine


def analyze_stock_performance(start_date, end_date):
    # --- 数据库配置 ---
    try:
        engine = create_engine(
            "mysql+pymysql://yangfan@8.159.129.156:10836/level?charset=utf8mb4",
            connect_args={
                "password": "aA%8023321088"
            },
            pool_pre_ping=True
        )

        # 1. 获取涨停当日股票及价格
        limit_query = f"""
        SELECT l.ts_code, l.name, d.open AS limit_day_open, d.high AS limit_day_high
        FROM stock_limit_data l
        LEFT JOIN stock_daily d ON l.ts_code = d.ts_code AND l.trade_date = d.trade_date
        WHERE l.trade_date = '{start_date}' AND l.status = 1
        """
        df_limit = pd.read_sql(limit_query, engine)

        if df_limit.empty:
            print(f"提示：{start_date} 当日没有涨停记录。")
            return

        # 2. 获取行情区间数据
        codes = df_limit['ts_code'].unique().tolist()
        codes_tuple = f"('{codes[0]}')" if len(codes) == 1 else str(tuple(codes))
        daily_query = f"""
        SELECT ts_code, trade_date, high, low, close
        FROM stock_daily
        WHERE ts_code IN {codes_tuple} AND trade_date BETWEEN '{start_date}' AND '{end_date}'
        ORDER BY trade_date ASC
        """
        df_daily = pd.read_sql(daily_query, engine)
        df_daily['trade_date'] = pd.to_datetime(df_daily['trade_date'])

        # 3. 核心逻辑处理
        final_report = []
        for _, limit_row in df_limit.iterrows():
            code = limit_row['ts_code']
            stock_data = df_daily[df_daily['ts_code'] == code].copy()

            if stock_data.empty: continue

            # A. 找到区间内的最高价及其第一次出现的日期
            max_high = float(stock_data['high'].max())
            # 获取最高价所在的那一行（如果有多个相同最高价，取第一个）
            max_high_row = stock_data[stock_data['high'] == max_high].iloc[0]
            max_date = max_high_row['trade_date']

            # B. 寻找最高价日期之后的最低价
            post_max_data = stock_data[stock_data['trade_date'] > max_date]

            if not post_max_data.empty:
                min_low = float(post_max_data['low'].min())
                last_close = float(post_max_data.iloc[-1]['close'])

                price_diff = max_high - min_low
                drop_pct = (price_diff / max_high * 100)
                rebound_pct = ((last_close - min_low) / min_low * 100) if min_low > 0 else 0
            else:
                # 如果最高点出现在结束日当天，则没有“后期最低价”
                min_low = None
                price_diff = 0
                drop_pct = 0
                rebound_pct = 0

            final_report.append({
                "代码": code,
                "名称": limit_row['name'],
                "当日开盘": float(limit_row['limit_day_open'] or 0),
                "当日最高": float(limit_row['limit_day_high'] or 0),
                "区间最高": max_high,
                "最高日期": max_date.strftime('%Y-%m-%d'),
                "见顶后最低": min_low,
                "最大回落幅%": round(drop_pct, 2),
                "低位至期末波动%": round(rebound_pct, 2)
            })

        # 4. 排序与输出
        if final_report:
            result_df = pd.DataFrame(final_report)
            result_df = result_df.sort_values(by="最大回落幅%", ascending=False)

            print(f"\n分析区间: {start_date} 至 {end_date}")
            print("说明: '见顶后最低' 指最高价日期之后的最低点")
            print("=" * 120)
            print(result_df.to_string(index=False, na_rep='N/A'))
            print("=" * 120)

    except Exception as e:
        print(f"发生错误: {e}")


# ==========================================
# 设定日期
# ==========================================
if __name__ == "__main__":
    START_INPUT = '2024-09-24'
    END_INPUT = '2026-04-23'

    analyze_stock_performance(START_INPUT, END_INPUT)