#!/usr/bin/env python3
"""
小唯四维选股系统 — 宏观+基本面+技术+消息
========================================
综合四个维度筛选股票

维度说明:
1. 宏观面 — 美联储政策/人民币汇率/地缘风险
2. 基本面 — PE/PB/业绩增速/行业景气度
3. 技术面 — MA20趋势/成交量/均线排列
4. 消息面 — 政策利好/行业新闻/重大事件

评分规则:
- 宏观负面(-1)/中性(0)/正面(+1)
- 基本面低估(+1)/合理(0)/高估(-1)
- 技术面多头(+1)/空头(-1)/震荡(0)
- 消息面利好(+1)/利空(-1)/中性(0)
- 综合评分: 4分以上关注, 6分以上重点

用法:
  python3 stock_selector.py scan              # 扫描全市场
  python3 stock_selector.py analyze <代码>     # 分析单只股票
  python3 stock_selector.py watchlist         # 展示关注列表
"""

import json, sys, urllib.request, time
from datetime import datetime
from pathlib import Path

import numpy as np
import pandas as pd

OUTPUT = Path.home() / ".hermes" / "stock_backtest"
OUTPUT.mkdir(exist_ok=True)
FEISHU_WEBHOOK = "https://open.feishu.cn/open-apis/bot/v2/hook/446db983-e392-4d2c-bfb8-f9060e5df3ad"

# ===================== 飞书推送 =====================

def send_feishu(msg):
    payload = json.dumps({"msg_type": "text", "content": {"text": msg}}).encode()
    req = urllib.request.Request(FEISHU_WEBHOOK, data=payload,
        headers={"Content-Type": "application/json"})
    try:
        with urllib.request.urlopen(req, timeout=10):
            pass
    except Exception:
        pass


# ===================== 宏观面 — 汇率/利率/风险 =====================

def get_macro_score():
    """
    宏观面评分:
    考虑: 人民币汇率, 美联储政策, 地缘风险
    返回: -1(负面) / 0(中性) / +1(正面)
    """
    score = 0
    reasons = []

    # 人民币汇率
    try:
        url = "https://qt.gtimg.cn/q=usdcnh"
        text = urllib.request.urlopen(url, timeout=5).read().decode("gbk")
        parts = text.split("~")
        if len(parts) > 10:
            usdcnh = float(parts[3])
            if usdcnh > 7.3:
                score -= 1
                reasons.append(f"离岸人民币破7.3({usdcnh}) → 外资流出压力 ⬇️")
            else:
                score += 1
                reasons.append(f"人民币相对稳定({usdcnh}) → 外资流出压力小 ⬆️")
    except Exception:
        reasons.append("人民币汇率获取失败")

    # A股风险偏好(用沪深300判断)
    try:
        url = "https://web.ifzq.gtimg.cn/appstock/app/fqkline/get?_var=kline_dayqfq&param=sh510300,day,2026-07-01,2026-07-12,10,qfq"
        text = urllib.request.urlopen(url, timeout=5).read().decode("utf-8")
        data = json.loads(text.replace("kline_dayqfq=", "", 1))
        rows = data.get("data", {}).get("sh510300", {}).get("qfqday") or []
        if len(rows) >= 5:
            closes = [float(r[2]) for r in rows]
            if closes[-1] > closes[0]:
                score += 1
                reasons.append("沪深300近期上涨 → 市场风险偏好上升 ⬆️")
            else:
                score -= 1
                reasons.append("沪深300近期下跌 → 市场风险偏好下降 ⬇️")
    except Exception:
        pass

    # 平均
    avg_score = max(-1, min(1, score // max(1, len([r for r in reasons if r]))))
    return avg_score, reasons


# ===================== 基本面 — PE/PB/业绩 =====================

def get_fundamental_score(code):
    """
    基本面评分:
    基于估值(PE/PB)和近期趋势
    返回: -1(高估) / 0(合理) / +1(低估)
    """
    try:
        mc = f"sh{code}" if code.startswith("6") else f"sz{code}"
        url = f"https://qt.gtimg.cn/q={mc}"
        text = urllib.request.urlopen(url, timeout=5).read().decode("gbk")
        parts = text.split("~")
        if len(parts) < 40:
            return 0, [], {}

        pe = float(parts[39]) if parts[39] else 0
        pb = float(parts[46]) if parts[46] else 0
        price = float(parts[3])
        yclose = float(parts[4])

        score = 0
        reasons = []

        # PE判断
        if 0 < pe < 15:
            score += 1
            reasons.append(f"PE={pe:.1f}(历史低位) → 估值有支撑 ⬆️")
        elif 15 <= pe <= 30:
            reasons.append(f"PE={pe:.1f}(合理区间) → 无明显低估/高估")
        elif pe > 30:
            score -= 1
            reasons.append(f"PE={pe:.1f}(历史高位) → 估值偏高风险 ⬇️")
        elif pe <= 0:
            reasons.append(f"PE={pe}(亏损/无效) → 无法判断")

        # PB判断
        if 0 < pb < 3:
            score += 1
            reasons.append(f"PB={pb:.1f}(相对低估) → 净资产有支撑 ⬆️")
        elif pb >= 3:
            reasons.append(f"PB={pb:.1f}(相对高估)")

        # 近期价格位置
        try:
            url2 = f"https://web.ifzq.gtimg.cn/appstock/app/fqkline/get?_var=kline_dayqfq&param={mc},day,2026-01-01,2026-07-12,200,qfq"
            text2 = urllib.request.urlopen(url2, timeout=5).read().decode("utf-8")
            data2 = json.loads(text2.replace("kline_dayqfq=", "", 1))
            hist = data2.get("data", {}).get(mc, {}).get("qfqday") or []
            if len(hist) >= 60:
                highs = [float(r[3]) for r in hist]
                low = min(highs)
                high = max(highs)
                pos = (price - low) / (high - low) if high > low else 0.5
                if pos < 0.2:
                    score += 1
                    reasons.append(f"价格处于近半年低位({pos*100:.0f}%) → 相对安全边际 ⬆️")
                elif pos > 0.8:
                    score -= 1
                    reasons.append(f"价格处于近半年高位({pos*100:.0f}%) → 追高风险 ⬇️")
        except Exception:
            pass

        avg_score = max(-1, min(1, score))
        info = {"pe": pe, "pb": pb, "price": price, "yclose": yclose}
        return avg_score, reasons, info
    except Exception as e:
        return 0, [f"基本面数据获取失败({e})"], {}


# ===================== 技术面 — MA20趋势 =====================

def get_technical_score(code):
    """
    技术面评分:
    基于MA20均线状态
    返回: -1(空头) / 0(震荡) / +1(多头)
    """
    try:
        mc = f"sh{code}" if code.startswith("6") else f"sz{code}"
        end = datetime.now().strftime("%Y-%m-%d")
        url = f"https://web.ifzq.gtimg.cn/appstock/app/fqkline/get?_var=kline_dayqfq&param={mc},day,2025-07-01,{end},500,qfq"
        text = urllib.request.urlopen(url, timeout=8).read().decode("utf-8")
        data = json.loads(text.replace("kline_dayqfq=", "", 1))
        rows = data.get("data", {}).get(mc, {}).get("qfqday") or []
        if len(rows) < 30:
            return 0, ["数据不足"], {}

        closes = [float(r[2]) for r in rows]
        dates = [r[0] for r in rows]
        closes_s = pd.Series(closes)

        ma20 = closes_s.rolling(20).mean()
        ma60 = closes_s.rolling(60).mean()
        ma120 = closes_s.rolling(120).mean()

        last_close = closes[-1]
        last_ma20 = ma20.iloc[-1]
        last_ma60 = ma60.iloc[-1]
        last_ma120 = ma120.iloc[-1]

        # 均线排列
        if last_ma20 > last_ma60 > last_ma120:
            ma_arrangement = "多头排列"
            t_score = 1
        elif last_ma20 < last_ma60 < last_ma120:
            ma_arrangement = "空头排列"
            t_score = -1
        else:
            ma_arrangement = "均线混乱"
            t_score = 0

        # 价格与MA20关系
        pct_above = (last_close - last_ma20) / last_ma20 * 100

        # 近期趋势
        trend_20d = (closes[-1] - closes[-20]) / closes[-20] * 100 if len(closes) >= 20 else 0

        reasons = [
            f"{ma_arrangement}(价格{'' if last_close>last_ma20 else ''}{pct_above:+.1f}%vsMA20)",
            f"近20日涨跌: {trend_20d:+.1f}%",
        ]

        info = {
            "price": last_close,
            "ma20": last_ma20,
            "ma60": last_ma60,
            "ma120": last_ma120,
            "ma_arrangement": ma_arrangement,
            "pct_above_ma20": pct_above,
            "trend_20d": trend_20d,
        }
        return t_score, reasons, info
    except Exception as e:
        return 0, [f"技术面数据获取失败({e})"], {}


# ===================== 消息面 — 行业/政策 =====================

def get_stock_news(code, name, max_news=3):
    """获取个股最新新闻（东方财富）"""
    try:
        url = f"https://np-anotice-stock.eastmoney.com/api/security/ann?sr=-1&page_size={max_news}&page_index=1&ann_type=SZA&stock_list={code}"
        import subprocess, shlex, os, json
        env = dict(os.environ)
        for k in ["http_proxy", "https_proxy", "HTTP_PROXY", "HTTPS_PROXY"]:
            env.pop(k, None)
        cmd = f"curl -s --max-time 6 --compressed {shlex.quote(url)}"
        r = subprocess.run(cmd, shell=True, capture_output=True, timeout=8, env=env)
        data = json.loads(r.stdout.decode("utf-8", errors="ignore"))
        notices = data.get("data", {}).get("list", [])
        news = []
        for n in notices[:max_news]:
            title = n.get("title", "")[:50]
            notice_date = n.get("notice_date", "")[:10]
            if title:
                news.append(f"· {notice_date} {title}")
        return news if news else [f"近{max_news}日无重大公告"]
    except Exception:
        return ["新闻获取失败"]


def get_sector_news(industry, max_news=2):
    """获取行业相关新闻 — 回退到指数情绪代理（行业API不可用时）"""
    try:
        # 先尝试东财快讯
        keyword_map = {"白酒": "白酒", "银行": "银行", "保险": "保险", "新能源": "新能源"}
        keyword = keyword_map.get(industry, industry) or industry or ""
        base = "https://search-api-web.eastmoney.com/search/jsonp"
        param = ('{"uid":"","keyword":"' + keyword + '","type":["cmsArticleListNew"],'
                 '"client":"web","clientVersion":"curr","clientType":"web",'
                 '"param":{"cmsArticleListNew":{"pageIndex":1,"pageSize":' + str(max_news) + '}}}')
        url = base + "?param=" + param
        env = dict(os.environ)
        for k in ["http_proxy", "https_proxy", "HTTP_PROXY", "HTTPS_PROXY"]:
            env.pop(k, None)
        cmd = ["curl", "-s", "--max-time", "6",
               "-H", "Referer: https://so.eastmoney.com/",
               "-H", "User-Agent: Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 Chrome/120 Safari/537.36",
               url]
        r = subprocess.run(cmd, capture_output=True, timeout=8, env=env)
        raw = r.stdout.decode("utf-8", errors="ignore")
        m = re.search(r'\(\[.*\]\)', raw, re.DOTALL)
        if m:
            items = json.loads(m.group(1))
            news = [f"· {it.get('title','')[:40]}" for it in items[:max_news] if it.get('title')]
            if news:
                return news
    except Exception:
        pass
    # 回退：使用指数涨跌作为行业情绪代理
    try:
        sector_codes = {"白酒": "sh000858", "银行": "sh000001", "保险": "sh601318", "新能源": "sz399808"}
        code = sector_codes.get(industry, "sh000001")
        mc = code[:2] + code[2:]
        url = f"https://qt.gtimg.cn/q={mc}"
        text = get_url_gbk(url)
        parts = text.split("~")
        if len(parts) > 5:
            p = float(parts[3])
            y = float(parts[4])
            chg = (p-y)/y*100
            sentiment = "📈 行业指数上涨" if chg > 0 else "📉 行业指数下跌"
            return [f"· {sentiment} ({chg:+.2f}%) — {parts[1]}"]
    except Exception:
        pass
    return ["· 行业新闻获取失败（回退到静态规则）"]


def get_sentiment_score_from_news(code, name="", industry=""):
    """消息面评分 — 基于真实新闻 + 静态规则回退"""
    import time
    news = get_stock_news(code, name)
    sector_news = get_sector_news(industry) if industry else []
    score = 0
    reasons = []

    # 真实新闻分析
    all_news = news + sector_news
    for item in all_news:
        text = item.lower()
        if any(k in text for k in ["违规", "调查", "处罚", "减持", "业绩预亏", "暴雷"]):
            score -= 1
            reasons.append(f"利空公告: {item[:30]}")
        elif any(k in text for k in ["回购", "增持", "业绩预增", "中标", "合作", "突破"]):
            score += 1
            reasons.append(f"利好公告: {item[:30]}")

    # 静态规则回退（无新闻时）
    if not reasons:
        if name in ["五粮液", "贵州茅台", "泸州老窖", "洋河股份"]:
            score -= 1; reasons.append("白酒消费降级压制 ⬇️")
            score += 1; reasons.append("估值低位有支撑 ⬆️")
        elif name in ["平安银行", "招商银行"]:
            score += 1; reasons.append("高股息防御 ⬆️")
            score += 1; reasons.append("低估值稳健 ⬆️")
        elif "宁德" in name:
            score -= 1; reasons.append("新能源产能过剩担忧 ⬇️")
            score += 1; reasons.append("行业龙头壁垒 ⬆️")
        else:
            reasons.append("消息面无明显驱动")

    avg_score = max(-1, min(1, score))
    return avg_score, reasons, news, sector_news


# ===================== 综合评分 =====================

def analyze_stock(code, name=""):
    """对单只股票进行四维评分"""
    print(f"\n{'='*60}")
    print(f"四维选股分析: {name}({code})")
    print(f"{'='*60}")

    # 1. 宏观面
    m_score, m_reasons = get_macro_score()
    print(f"[宏观面] 评分: {m_score:+d}")
    for r in m_reasons:
        print(f"  - {r}")

    # 2. 基本面
    f_score, f_reasons, f_info = get_fundamental_score(code)
    print(f"[基本面] 评分: {f_score:+d}")
    for r in f_reasons:
        print(f"  - {r}")

    # 3. 技术面
    t_score, t_reasons, t_info = get_technical_score(code)
    print(f"[技术面] 评分: {t_score:+d}")
    for r in t_reasons:
        print(f"  - {r}")

    # 4. 消息面
    industry_map = {"000858": "白酒", "600519": "白酒", "000568": "白酒", "002304": "白酒",
                    "600036": "银行", "601318": "保险", "000001": "银行", "300750": "新能源", "510300": ""}
    industry = industry_map.get(code, "")
    s_score, s_reasons, stock_news, sector_news = get_sentiment_score_from_news(code, name, industry)
    print(f"[消息面] 评分: {s_score:+d}")
    for r in s_reasons:
        print(f"  - {r}")
    if stock_news:
        for n in stock_news[:2]:
            print(f"  📰 {n}")
    if sector_news:
        for n in sector_news[:1]:
            print(f"  📢 {n}")

    # ===========================================================
    # 组合评分逻辑（修正简单相加的缺陷）
    # ===========================================================
    # 规则1: 技术面空头 + 基本面低估 = 买入机会(逆向)
    # 规则2: 技术面空头 + 基本面高估 = 危险加倍(双杀)
    # 规则3: 宏观负面时，技术面多头无法持续
    # 规则4: 综合评分考虑方向匹配

    f_strong = f_score >= 1   # 基本面好(低估)
    f_weak   = f_score <= -1  # 基本面差(高估)
    t_bear   = t_score <= -1  # 技术面空头(超跌)
    t_bull   = t_score >= 1   # 技术面多头
    m_bad    = m_score <= -1  # 宏观差

    # 逆向机会: 技术面空头+基本面好 = 低估买入机会
    if t_bear and f_strong:
        adjusted_total = 2  # 视为关注机会
        combo_reason = "逆向机会: 技术超跌+基本面低估"
    # 双杀: 技术差+基本面也差
    elif t_bear and f_weak:
        adjusted_total = -2
        combo_reason = "双杀风险: 技术超跌+基本面高估"
    # 技术多头+基本面好 + 宏观不差
    elif t_bull and f_strong and not m_bad:
        adjusted_total = 3
        combo_reason = "共振: 技术+基本面+宏观同向"
    # 强势股(宏观好时)
    elif t_bull and m_score >= 0:
        adjusted_total = 2
        combo_reason = "技术多头，宏观中性支撑"
    # 常规计算
    else:
        adjusted_total = m_score + f_score + t_score + s_score
        combo_reason = "常规评分"

    total = adjusted_total
    max_score = 4

    print(f"\n组合逻辑: {combo_reason}")
    print(f"\n{'='*60}")
    print(f"综合评分: {total:+d} / {max_score} ({(total/max_score*100):.0f}%)")
    if total >= 6:
        verdict = "⭐⭐⭐ 重点关注 — 四维共振，强烈看多"
    elif total >= 4:
        verdict = "⭐⭐ 关注 — 多维度偏多"
    elif total >= 2:
        verdict = "⭐ 关注 — 逆向机会或温和看多"
    elif total >= 0:
        verdict = "⚠️ 观察 — 方向不明"
    else:
        verdict = "❌ 不碰 — 双杀风险"
    print(f"结论: {verdict}")
    print(f"{'='*60}")

    # 保存
    result = {
        "code": code, "name": name, "date": datetime.now().strftime("%Y-%m-%d %H:%M"),
        "scores": {"macro": m_score, "fundamental": f_score, "technical": t_score, "sentiment": s_score},
        "total": total,
        "verdict": verdict,
        "macro_reasons": m_reasons,
        "fundamental_reasons": f_reasons,
        "fundamental_info": f_info,
        "technical_reasons": t_reasons,
        "technical_info": t_info,
        "sentiment_reasons": s_reasons,
    }
    result_file = OUTPUT / f"四维_{code}.json"
    with open(result_file, "w") as f:
        json.dump(result, f, ensure_ascii=False, indent=2)
    print(f"数据存: {result_file}")

    return result


def scan_watchlist():
    """扫描关注列表"""
    stocks = [
        ("000858", "五粮液"),
        ("600519", "贵州茅台"),
        ("000001", "平安银行"),
        ("300750", "宁德时代"),
        ("510300", "沪深300ETF"),
    ]

    print(f"\n{'='*70}")
    print(f"{'股票':<12} {'宏观':>5} {'基本面':>6} {'技术面':>6} {'消息面':>6} {'综合':>5} {'结论'}")
    print(f"{'-'*70}")

    results = []
    for code, name in stocks:
        r = analyze_stock(code, name)
        results.append(r)

        sc = r["scores"]
        ts = r["total"]
        print(f"{name:<12} {sc['macro']:>+4} {sc['fundamental']:>+5} {sc['technical']:>+5} {sc['sentiment']:>+5} {ts:>+4}  {r['verdict'].split(' ')[0]}")

        time.sleep(0.5)  # 避免请求过快

    print(f"{'='*70}")

    # 汇总
    print(f"\n📊 宏观面: {'负面 ⬇️' if results[0]['scores']['macro'] < 0 else '正面 ⬆️' if results[0]['scores']['macro'] > 0 else '中性'}")
    good = [r for r in results if r["total"] >= 2]
    print(f"可选股票({len(good)}只): {[r['name'] for r in good]}")

    msg = f"""📊 四维选股扫描

宏观面: {'负面 ⬇️' if results[0]['scores']['macro'] < 0 else '正面 ⬆️'}

| 股票 | 宏观 | 基本面 | 技术 | 消息 | 综合 |
|------|------|--------|------|------|------|
"""
    for r in results:
        sc = r["scores"]
        msg += f"| {r['name']} | {sc['macro']:>+2} | {sc['fundamental']:>+2} | {sc['technical']:>+2} | {sc['sentiment']:>+2} | **{r['total']:>+2}** |\n"

    good = [r for r in results if r["total"] >= 2]
    msg += f"\n可选股票({len(good)}): "
    msg += " / ".join([r["name"] for r in good]) if good else "无"

    msg += f"\n\n生成: {datetime.now().strftime('%Y-%m-%d %H:%M')}"
    send_feishu(msg)


if __name__ == "__main__":
    if "--scan" in sys.argv or "scan" in sys.argv:
        scan_watchlist()
    elif "--watchlist" in sys.argv:
        scan_watchlist()
    elif len(sys.argv) >= 3 and sys.argv[1] == "analyze":
        code = sys.argv[2]
        name = sys.argv[3] if len(sys.argv) > 3 else code
        analyze_stock(code, name)
    elif len(sys.argv) >= 2 and sys.argv[1] not in ["--scan", "--watchlist"]:
        code = sys.argv[1]
        name = sys.argv[2] if len(sys.argv) > 2 else code
        analyze_stock(code, name)
    else:
        print("用法:")
        print("  python3 stock_selector.py --scan                    # 扫描关注列表")
        print("  python3 stock_selector.py analyze <代码> [名称]     # 分析单只")
        print("  python3 stock_selector.py 000858 五粮液             # 同上简写")