MoFin/mo_alphasift_bridge.py

#!/usr/bin/env python3
"""
mo_alphasift_bridge.py — AlphaSift 多策略并行选股 → MoFin 自选池

支持同时跑多个策略，合并去重后写入自选池。
默认三策略: balanced_alpha + dual_low + quality_value

用法:
  python3 mo_alphasift_bridge.py                              # 默认三策略
  python3 mo_alphasift_bridge.py --strategy dual_low          # 单策略
  python3 mo_alphasift_bridge.py --dry-run                     # 只看不写
  python3 mo_alphasift_bridge.py --strategy list               # 列出所有策略
"""

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

DSA_API = "http://127.0.0.1:8001"
MOFIN_DATA = Path("/home/hmo/web-dashboard/data")
WATCHLIST_PATH = MOFIN_DATA / "watchlist.json"
PORTFOLIO_PATH = MOFIN_DATA / "portfolio.json"

DEFAULT_STRATEGIES = "balanced_alpha,dual_low,quality_value"
DEFAULT_MARKET = "cn"
DEFAULT_MAX = 15
MIN_SCORE = 5
MAX_ADD = 5
POLL_INTERVAL = 5
POLL_TIMEOUT = 300


def load_json(path):
    try: return json.loads(path.read_text(encoding="utf-8"))
    except: return None

def save_json(path, data):
    path.parent.mkdir(parents=True, exist_ok=True)
    path.write_text(json.dumps(data, ensure_ascii=False, indent=2), encoding="utf-8")

def api(endpoint, method="GET", body=None):
    url = f"{DSA_API}{endpoint}"
    data = json.dumps(body).encode() if body else None
    req = urllib.request.Request(url, data=data, method=method,
                                  headers={"Content-Type": "application/json"})
    try:
        with urllib.request.urlopen(req, timeout=30) as r:
            return json.loads(r.read())
    except Exception as e:
        print(f"  API错误: {e}")
        return None

def get_existing_codes():
    codes = set()
    for path in [WATCHLIST_PATH, PORTFOLIO_PATH]:
        data = load_json(path)
        if not data: continue
        key = "stocks" if "watchlist" in str(path) else "holdings"
        for item in data.get(key, []):
            c = str(item.get("code", "")).strip()
            if c: codes.add(c)
    return codes


def run_one_strategy(strategy, market, max_results):
    """跑单个策略，返回候选股列表"""
    print(f"\n{'='*50}")
    print(f"策略: {strategy}")
    print(f"{'='*50}")
    
    task = api("/api/v1/alphasift/screen/tasks", "POST", {
        "strategy": strategy, "market": market, "max_results": max_results
    })
    if not task or not task.get("task_id"):
        print(f"  FAIL: 提交失败")
        return []
    
    task_id = task["task_id"]
    print(f"  任务: {task_id[:12]}...", flush=True)
    
    waited = 0
    while waited < POLL_TIMEOUT:
        time.sleep(POLL_INTERVAL)
        waited += POLL_INTERVAL
        status = api(f"/api/v1/alphasift/screen/tasks/{task_id}")
        if not status: continue
        s = status.get("status", "")
        if s == "completed":
            print(f"  完成 ({waited}s)")
            result = status.get("result", {})
            candidates = result.get("candidates", [])
            print(f"  候选: {len(candidates)} 只")
            if candidates:
                for c in candidates[:3]:
                    print(f"    {c.get('code','?')} {c.get('name',c.get('title','?'))} 评分{c.get('score','?'):.1f}")
            return candidates
        elif s == "failed":
            print(f"  FAIL: {status.get('error','')}")
            return []
        else:
            if waited % 60 == 0:
                print(f"  ...{s} ({status.get('progress',0)}%)", flush=True)
    
    print(f"  FAIL: 超时")
    return []


def run_all(strategies_str, market, max_results, dry_run=False):
    """多策略并行 → 合并去重 → MoFin 自选池"""
    strategies = [s.strip() for s in strategies_str.split(",") if s.strip()]
    now = datetime.now()
    date_str = now.strftime("%Y-%m-%d")
    time_str = now.strftime("%Y-%m-%d %H:%M")
    
    print(f"AlphaSift 多策略选股: {', '.join(strategies)}")
    print(f"开始: {time_str}")
    
    # 逐个跑策略，汇总
    all_candidates = []
    seen = set()
    for strategy in strategies:
        candidates = run_one_strategy(strategy, market, max_results)
        for c in candidates:
            code = str(c.get("code", "")).strip()
            if code in seen: continue
            seen.add(code)
            c["_strategy"] = strategy
            all_candidates.append(c)
    
    if not all_candidates:
        print("\n无候选股")
        return
    
    print(f"\n汇总: {len(all_candidates)} 只候选股 (去重后)")
    for s in strategies:
        cnt = sum(1 for c in all_candidates if c.get("_strategy") == s)
        print(f"  {s}: {cnt} 只")
    
    # 过滤
    existing = get_existing_codes()
    new_stocks = []
    skipped_score = 0
    skipped_dup = 0
    
    for c in all_candidates:
        code = str(c.get("code", "")).strip()
        score = c.get("score", 0) or c.get("llm_score", 0) or 0
        if score < MIN_SCORE:
            skipped_score += 1; continue
        if code in existing:
            skipped_dup += 1; continue
        
        name = c.get("name", "") or c.get("title", "") or code
        reason = c.get("reason", "") or c.get("llm_thesis", "")
        src = c.get("_strategy", "unknown")
        
        factors = c.get("factor_scores", {})
        factor_note = ", ".join(f"{k}={v:.0f}" for k,v in list(factors.items())[:3]) if factors else ""
        
        notes = f"AlphaSift/{src} 评分{score:.0f}"
        if factor_note: notes += f" [{factor_note}]"
        if reason: notes += f" | {reason[:120]}"
        
        new_stocks.append({
            "code": code,
            "name": name,
            "price": c.get("price", 0),
            "source": "alpha_sift",
            "source_detail": {
                "strategy": src,
                "strategies_run": strategies,
                "score": score,
                "factor_scores": factors,
                "date": date_str,
                "reason": reason[:300],
            },
            "notes": notes,
            "added_at": time_str,
            "added_by": "AlphaSift",
            "analysis": {},
        })
        existing.add(code)
    
    print(f"\n过滤: {len(new_stocks)} 新标的 (评分不足{skipped_score} + 重复{skipped_dup})")
    
    if not new_stocks:
        print("无符合条件的新标的")
        return
    
    new_stocks = new_stocks[:MAX_ADD]
    
    print(f"\n新增 {len(new_stocks)} 只到自选池:")
    for s in new_stocks:
        sd = s["source_detail"]
        print(f"  {s['code']} {s['name']} ({sd['strategy']} 评分{sd['score']:.0f})")
    
    if dry_run:
        print("\n[DRY RUN] 未写入")
        return
    
    # 写入
    wl = load_json(WATCHLIST_PATH) or {"stocks": []}
    wl["stocks"].extend(new_stocks)
    wl["updated_at"] = time_str
    save_json(WATCHLIST_PATH, wl)
    print(f"\n已写入 {WATCHLIST_PATH}")
    
    # 策略生成
    print("\n调用 regenerate_all()...")
    try:
        sys.path.insert(0, str(MOFIN_DATA.parent))
        from strategy_lifecycle import regenerate_all
        r = regenerate_all(stdout=True)
        if r: print(f"完成: {r.get('ok',0)}/{r.get('total',0)} 只策略已生成")
    except Exception as e:
        print(f"WARN: {e}")


def list_strategies():
    r = api("/api/v1/alphasift/strategies")
    if r and r.get("strategies"):
        for s in r["strategies"]:
            print(f"  {s['id']:22s} {s.get('name','?'):10s} {s.get('description','')[:60]}")


def main():
    global MIN_SCORE
    p = argparse.ArgumentParser(description="AlphaSift → MoFin")
    p.add_argument("--strategy", default=DEFAULT_STRATEGIES)
    p.add_argument("--market", default=DEFAULT_MARKET)
    p.add_argument("--max", type=int, default=DEFAULT_MAX)
    p.add_argument("--min-score", type=int, default=MIN_SCORE)
    p.add_argument("--dry-run", action="store_true")
    args = p.parse_args()
    MIN_SCORE = args.min_score
    if args.strategy == "list": list_strategies()
    else: run_all(args.strategy, args.market, args.max, args.dry_run)


if __name__ == "__main__":
    main()