feat: cninfo行业富集器——批量补全stock_sectors行业映射+回填sector_context

2026-07-22 13:04:43 +08:00
parent 5d8c7476ff
commit cdfc7346f1
1 changed files with 60 additions and 0 deletions
@@ -0,0 +1,60 @@
+#!/usr/bin/env python3
+"""sector_enrich_cninfo.py — 用巨潮 cninfo 批量补全行业/主营业务映射
+覆盖全部 active holding_strategies。限速 0.3s/只。
+"""
+import sqlite3, time, sys
+import akshare as ak
+
+DB = "/home/hmo/MoFin/data/mofin.db"
+
+def main():
+    conn = sqlite3.connect(DB)
+    codes = [r[0] for r in conn.execute(
+        "SELECT DISTINCT code FROM holding_strategies WHERE status='active'")]
+    have = {r[0] for r in conn.execute("SELECT code FROM stock_sectors")}
+    todo = [c for c in codes if c not in have]
+    print(f"总 {len(codes)} 只，缺 {len(todo)} 只")
+
+    ok = fail = 0
+    for i, code in enumerate(todo, 1):
+        # 港股 (5位 0/1 开头) cninfo 不支持
+        if len(str(code)) == 5 and str(code)[0] in "01":
+            continue
+        try:
+            df = ak.stock_profile_cninfo(symbol=str(code))
+            if df is not None and len(df):
+                row = df.iloc[0].to_dict()
+                sector = str(row.get("所属行业") or "").strip()
+                business = str(row.get("主营业务") or "").strip()[:200]
+                if sector:
+                    conn.execute(
+                        "INSERT OR REPLACE INTO stock_sectors (code, sector_name, source, updated_at) "
+                        "VALUES (?, ?, 'cninfo', datetime('now','localtime'))",
+                        (code, sector))
+                    if business:
+                        # business 存到 sector_name 扩展？表只有 sector_name，把 business 塞进 source? 不，加列太多。
+                        # sector_name 保留行业，business 记到 holding_strategies.sector_context（如果为空/被污染）
+                        conn.execute(
+                            "UPDATE holding_strategies SET sector_context=? WHERE code=? AND status='active' "
+                            "AND (sector_context IS NULL OR sector_context='' OR sector_context LIKE '大盘上涨比%')",
+                            (f"行业{sector} | {business[:60]}", code))
+                    ok += 1
+                else:
+                    fail += 1
+            else:
+                fail += 1
+        except Exception as e:
+            fail += 1
+            if fail <= 3:
+                print(f"  {code} 失败: {str(e)[:80]}")
+        if i % 20 == 0:
+            conn.commit()
+            print(f"  进度 {i}/{len(todo)} ok={ok} fail={fail}")
+        time.sleep(0.3)
+    conn.commit()
+    total = conn.execute("SELECT COUNT(*) FROM stock_sectors").fetchone()[0]
+    print(f"完成: 新增 {ok}, 失败 {fail}, stock_sectors 总数 {total}")
+    conn.close()
+
+if __name__ == "__main__":
+    main()