#!/usr/bin/env python3
"""step43_negative_factor.py — 大回撤案例负面因子提取
方法：对每个信号模拟交易（信号日买→止损/止盈/40日），
分组：亏损组(ret<0) vs 盈利组(ret>0)，对比信号日12维特征+市场气氛，
提取亏损交易共有的负面因子。
"""
import numpy as np
import pandas as pd
import sqlite3

print("=== 加载 ===", flush=True)
panel = pd.read_pickle("/tmp/panel_12d.pkl")
panel = panel.sort_values(["code", "date"]).reset_index(drop=True)
panel["_key"] = panel["code"] + "_" + panel["date"]
pos_map = {k: i for i, k in enumerate(panel["_key"])}
print("面板:", len(panel), flush=True)

# 大盘/气氛指标
conn = sqlite3.connect("file:/home/hmo/MoFin/data/mofin.db?mode=ro", uri=True)
idx_df = pd.read_sql("SELECT date, close, high FROM stock_daily WHERE code='sh000001' ORDER BY date", conn)
idx_df["date"] = idx_df["date"].astype(str)
idx_df["mkt_ret5"] = idx_df["close"].pct_change(5) * 100
idx_df["hi60"] = idx_df["high"].rolling(60).max()
idx_df["mkt_dd60"] = (idx_df["close"] / idx_df["hi60"] - 1) * 100
close_arr = idx_df["close"].values
down_days, streak, prev = [], 0, None
for c in close_arr:
    streak = streak + 1 if (prev is not None and c < prev) else 0
    down_days.append(streak)
    prev = c
idx_df["mkt_down_days"] = down_days
mkt_map = {}
for r in idx_df.itertuples():
    mkt_map[r.date] = (r.mkt_ret5, r.mkt_dd60, r.mkt_down_days)

news_cnt = pd.read_sql("SELECT substr(date,1,10) d, COUNT(*) c FROM stock_news GROUP BY d", conn)
news_cnt["d"] = news_cnt["d"].astype(str)
news_cnt["mkt_news5"] = news_cnt["c"].rolling(5, min_periods=1).mean()
news_map = dict(zip(news_cnt["d"], news_cnt["mkt_news5"]))
news_med = news_cnt["mkt_news5"].median()
print("新闻中位:", news_med, flush=True)

# 信号（step37 同款）
sig_cond = (
    (panel["mkt_rsi"] < 50) & (panel["mcap_q"] < 0.2) & (panel["pe_q"] < 0.2) &
    (panel["news3"] >= 1) & (panel["sec_ret20"] < 0) & (panel["bias60"] < -20)
)
panel["_dd60"] = panel["date"].map(lambda d: mkt_map[d][1] if d in mkt_map else np.nan)
sig_cond = sig_cond & (panel["_dd60"] <= -5)
cand = panel[sig_cond][["code", "date"]].copy()
cand = cand.sort_values(["code", "date"])
cand["prev"] = cand.groupby("code")["date"].shift(1)
cand["gap"] = (pd.to_datetime(cand["date"]) - pd.to_datetime(cand["prev"])).dt.days
cand = cand[(cand["prev"].isna()) | (cand["gap"] > 30)]
print("信号:", len(cand), flush=True)

# K线
codes = cand["code"].unique().tolist()
ph = ",".join("?" * len(codes))
df = pd.read_sql("SELECT code, date, close, high, low FROM stock_daily WHERE code IN ({}) ORDER BY code, date".format(ph), conn, params=codes)
df["date"] = df["date"].astype(str)
df["code"] = df["code"].astype(str).str.zfill(6)
df = df.sort_values(["code", "date"]).reset_index(drop=True)
df["_key"] = df["code"] + "_" + df["date"]
dpos = {k: i for i, k in enumerate(df["_key"])}
closes = df["close"].values
highs = df["high"].values
lows = df["low"].values
cand["kidx"] = (cand["code"] + "_" + cand["date"]).map(dpos)
cand = cand.dropna(subset=["kidx"]).copy()
cand["kidx"] = cand["kidx"].astype(int)
print("可定位:", len(cand), flush=True)

# 支撑压力（用于止损止盈）
import sys
sys.path.insert(0, "/tmp")
from sr_calculator import SRCalculator
sr = SRCalculator()
print("=== 计算支撑压力 + 模拟交易 ===", flush=True)
STOP_BUF = 0.05
HOLD = 40
rows = []
for i, r in enumerate(cand.itertuples()):
    if i % 500 == 0:
        print("  {} / {}".format(i, len(cand)), flush=True)
    bars_code = sr.get_bars(r.code)
    d_idx = bars_code.index[bars_code["date"] == r.date]
    if len(d_idx) == 0:
        continue
    sr_full = sr.sr_full(r.code, d_idx[0])
    pv = sr_full["pivot"]
    chip = sr_full["chip"]
    if not pv:
        continue
    sig_close = closes[r.kidx]
    support = pv["s2"]
    if chip and chip["chip_ss"] < sig_close:
        support = max(pv["s2"], chip["chip_ss"])
    resist = pv["r2"]
    if chip and chip["chip_sr"] > sig_close:
        resist = min(pv["r2"], chip["chip_sr"])
    if support >= sig_close or resist <= sig_close:
        continue
    # 模拟：信号日买，支撑下5%止损/压力止盈/40日
    stop = support * (1 - STOP_BUF)
    tp = resist
    ret = None
    reason = None
    for j in range(r.kidx+1, min(r.kidx+HOLD+1, len(closes))):
        hi, lo = highs[j], lows[j]
        if hi >= tp:
            ret = (tp / sig_close - 1) * 100
            reason = "tp"
            break
        if lo <= stop:
            ret = (lo / sig_close - 1) * 100
            reason = "stop"
            break
    if ret is None:
        ret = (closes[min(r.kidx+HOLD, len(closes)-1)] / sig_close - 1) * 100
        reason = "hold40"
    # 信号日特征（12维 + 气氛）
    row = {"code": r.code, "date": r.date, "ret": ret, "reason": reason,
           "win": 1 if ret > 0 else 0, "ret_abs": abs(ret)}
    feat = panel.iloc[pos_map[r.code + "_" + r.date]]
    for col in ["mkt_rsi", "mkt_adx", "sec_ret20", "sec_above", "rsi", "bias60",
                "bias20", "dist_lo20", "ret1", "ret5", "ret20", "vol_ratio",
                "limit_up", "hi20_new", "news3", "flow1", "flow5",
                "mcap_q", "pe_q", "pb_q"]:
        row["sig_" + col] = feat[col]
    # 气氛
    mk = mkt_map.get(r.date, (np.nan, np.nan, np.nan))
    row["mkt_ret5"] = mk[0]
    row["mkt_dd60"] = mk[1]
    row["mkt_down_days"] = mk[2]
    row["mkt_news5"] = news_map.get(r.date, np.nan)
    rows.append(row)

tr = pd.DataFrame(rows)
print("\n交易:", len(tr), flush=True)
tr.to_csv("/tmp/step43_trades.csv", index=False)

# ── 负面因子提取：亏损 vs 盈利 ──
print("\n=== 亏损 vs 盈利 特征对比 ===", flush=True)
win = tr[tr["win"] == 1]
lose = tr[tr["win"] == 0]
print("盈利: {} ({:.1f}%) 亏损: {} ({:.1f}%)".format(
    len(win), len(win)/len(tr)*100, len(lose), len(lose)/len(tr)*100), flush=True)
print("盈利 avg={:.2f}% 亏损 avg={:.2f}%".format(
    win["ret"].mean(), lose["ret"].mean()), flush=True)

out = ["# 大回撤负面因子提取\n",
       "- 盈利 {} 笔 avg{:.2f}% / 亏损 {} 笔 avg{:.2f}%".format(
           len(win), win["ret"].mean(), len(lose), lose["ret"].mean()),
       "- 对比信号日特征：亏损组显著偏离盈利组的 = 负面因子\n",
       "| 特征 | 亏损组 | 盈利组 | 差异 | 方向 |",
       "|---|---:|---:|---:|---|"]

feat_cols = [c for c in tr.columns if c.startswith("sig_")] + \
            ["mkt_ret5", "mkt_dd60", "mkt_down_days", "mkt_news5"]
for col in feat_cols:
    w = win[col].dropna()
    l = lose[col].dropna()
    if len(w) < 50 or len(l) < 50:
        continue
    wm, lm = w.mean(), l.mean()
    diff = lm - wm
    rel = abs(diff) / max(abs(wm), 1e-9)
    if rel > 0.05:
        out.append("| {} | {:.3f} | {:.3f} | {:.3f} | {} |".format(
            col, lm, wm, diff, "↑亏高" if diff > 0 else "↓亏低"))

# 大亏损（<-10%）特征
out.append("\n## 大亏损（ret<-10%）vs 盈利 特征\n")
big_lose = tr[tr["ret"] < -10]
out.append("- 大亏损 {} 笔 ({}%)".format(len(big_lose), len(big_lose)/len(tr)*100))
out.append("| 特征 | 大亏损组 | 盈利组 | 差异 |")
out.append("|---|---:|---:|---:|")
for col in feat_cols:
    b = big_lose[col].dropna()
    w = win[col].dropna()
    if len(b) < 20 or len(w) < 50:
        continue
    bm, wm = b.mean(), w.mean()
    diff = bm - wm
    rel = abs(diff) / max(abs(wm), 1e-9)
    if rel > 0.05:
        out.append("| {} | {:.3f} | {:.3f} | {:.3f} |".format(col, bm, wm, diff))

# 卖出原因分布
out.append("\n## 卖出原因 × 盈亏\n")
out.append("| 原因 | 笔数 | 平均收益 | 胜率 |")
out.append("|---|---:|---:|---:|")
for rs in tr["reason"].unique():
    sub = tr[tr["reason"] == rs]
    out.append("| {} | {} | {:.2f}% | {:.1f}% |".format(
        rs, len(sub), sub["ret"].mean(), (sub["ret"]>0).mean()*100))

with open("/tmp/step43_negative_report.md", "w", encoding="utf-8") as f:
    f.write("\n".join(out))
print("\n报告已写 /tmp/step43_negative_report.md", flush=True)
print("=== 完成 ===", flush=True)