perf: 6项性能与正确性优化

1. get_stock_name: 延迟批量保存缓存(60秒去抖)，避免每次新缓存都写文件 2. get_stock_name: 优先从DB stock_realtime_price查name，未命中再调腾讯API 3. compute_comprehensive_score: 单股模式也使用_fund_flow_cache(30分钟TTL) 4. _get_kline_from_local_db: 移除conn.autocommit=True，避免污染连接池事务模式 5. deep_analyze: K线天数从180改为120，与实时检测统一 6. detect_all_signals: 已有指标列时跳过重复calc_all_indicators
2026-07-22 08:00:27 +08:00
parent 037d9cd26c
commit cf7341fc63
5 changed files with 60 additions and 19 deletions
@@ -340,7 +340,9 @@ def detect_all_signals(df, lookback=5):
        if not np.issubdtype(df[col].dtype, np.floating):
            df[col] = pd.to_numeric(df[col], errors='coerce').fillna(0).astype(np.float64)

-    df = calc_all_indicators(df)
+    # 若已有指标列则跳过重复计算（deep_analyze 等场景预计算过）
+    if 'dif' not in df.columns or 'ma5' not in df.columns:
+        df = calc_all_indicators(df)

    all_signals = []
    all_signals.extend(detect_main_rising_wave(df, lookback))