提交

backend/app/data/company.py

@@ -9,7 +9,7 @@ from __future__ import annotations
 
 import asyncio
 import time
-from datetime import datetime, timedelta, timezone
+from datetime import datetime
 
 from sqlalchemy import select
 from sqlalchemy.dialects.postgresql import insert as pg_insert
@@ -18,12 +18,10 @@ from sqlalchemy.ext.asyncio import AsyncSession
 from ..config import settings
 from ..db import async_session
 from ..models import StockCompany
+from .sync_utils import call_retry, f_clean, fresh, get_pro_lazy, s_clean, utcnow
 
 _REFRESH_DAYS = 30
 
-# 频率超限特征（等待 62s 重试一次；与 screener.market_sync / data.etf_sync._call_retry 同款语义）
-_RATE_MARKS = ("频率超限", "每分钟")
-
 # 显式列出全部字段：introduction/office/main_business/business_scope 文档标注默认不显示，
 # 不传 fields 时 tushare 不返回这四列（实测 000001.SZ）
 _FIELDS = (
@@ -32,87 +30,38 @@ _FIELDS = (
     "employees,main_business,business_scope"
 )
 
-_pro = None  # 惰性单例（get_pro 每次都 ts.set_token 写文件，没必要重复）
-
-
-def _get_pro():
-    if not settings.tushare_token:
-        raise RuntimeError("未配置 TUSHARE_TOKEN，无法拉取公司简介（backend/.env）")
-    global _pro
-    if _pro is None:
-        from .tushare_provider import get_pro
-
-        _pro = get_pro()
-    return _pro
-
-
-def _call_retry(fn, *args, **kwargs):
-    """同步调用 tushare 接口；「每分钟」级频率超限等 62s 重试一次。"""
-    try:
-        return fn(*args, **kwargs)
-    except Exception as e:  # noqa: BLE001
-        msg = str(e)
-        if any(m in msg for m in _RATE_MARKS) and "小时" not in msg:
-            time.sleep(62)
-            return fn(*args, **kwargs)
-        raise
-
-
-def _utcnow() -> datetime:
-    return datetime.now(timezone.utc).replace(tzinfo=None)
-
-
-def _fresh(updated_at: datetime | None) -> bool:
-    return updated_at is not None and updated_at >= _utcnow() - timedelta(days=_REFRESH_DAYS)
-
-
-def _s(v) -> str | None:
-    """pandas NaN / 空串 / None -> None，其余 strip。"""
-    if v is None or (isinstance(v, float) and v != v):
-        return None
-    s = str(v).strip()
-    return s or None
-
-
-def _f(v) -> float | None:
-    try:
-        f = float(v)
-    except (TypeError, ValueError):
-        return None
-    return None if f != f else f  # NaN -> None
-
 
 def _i(v) -> int | None:
-    f = _f(v)
+    f = f_clean(v)
     return None if f is None else int(f)
 
 
 def fetch_company_sync(ts_code: str) -> dict | None:
     """同步拉单只公司简介（需在 to_thread 里跑）；返回行 dict，无此股返回 None。"""
     time.sleep(settings.screener_sync_interval)
-    df = _call_retry(_get_pro().stock_company, ts_code=ts_code, fields=_FIELDS)
+    df = call_retry(get_pro_lazy().stock_company, ts_code=ts_code, fields=_FIELDS)
     if df is None or df.empty:
         return None
     r = df.iloc[0]
     return {
         "ts_code": ts_code,
-        "com_name": _s(r.get("com_name")),
-        "com_id": _s(r.get("com_id")),
-        "chairman": _s(r.get("chairman")),
-        "manager": _s(r.get("manager")),
-        "secretary": _s(r.get("secretary")),
-        "reg_capital": _f(r.get("reg_capital")),
-        "setup_date": _s(r.get("setup_date")),
-        "province": _s(r.get("province")),
-        "city": _s(r.get("city")),
-        "introduction": _s(r.get("introduction")),
-        "website": _s(r.get("website")),
-        "email": _s(r.get("email")),
-        "office": _s(r.get("office")),
+        "com_name": s_clean(r.get("com_name")),
+        "com_id": s_clean(r.get("com_id")),
+        "chairman": s_clean(r.get("chairman")),
+        "manager": s_clean(r.get("manager")),
+        "secretary": s_clean(r.get("secretary")),
+        "reg_capital": f_clean(r.get("reg_capital")),
+        "setup_date": s_clean(r.get("setup_date")),
+        "province": s_clean(r.get("province")),
+        "city": s_clean(r.get("city")),
+        "introduction": s_clean(r.get("introduction")),
+        "website": s_clean(r.get("website")),
+        "email": s_clean(r.get("email")),
+        "office": s_clean(r.get("office")),
         "employees": _i(r.get("employees")),
-        "main_business": _s(r.get("main_business")),
-        "business_scope": _s(r.get("business_scope")),
-        "updated_at": _utcnow(),
+        "main_business": s_clean(r.get("main_business")),
+        "business_scope": s_clean(r.get("business_scope")),
+        "updated_at": utcnow(),
     }
 
 
@@ -151,7 +100,7 @@ async def get_company(session: AsyncSession, ts_code: str) -> dict | None:
     """
     row = (await session.execute(
         select(StockCompany).where(StockCompany.ts_code == ts_code))).scalar_one_or_none()
-    if row is not None and _fresh(row.updated_at):
+    if row is not None and fresh(row.updated_at, _REFRESH_DAYS):
         return _row_dict(row) if row.com_name is not None else None  # 墓碑 -> None
     # 释放请求会话持有的连接：后面可能隔着 1-2s 的 tushare 调用，别长占连接池。
     # 用 close() 而非 rollback()：rollback 会把会话身份映射里的实例全部 expire——
@@ -165,7 +114,7 @@ async def get_company(session: AsyncSession, ts_code: str) -> dict | None:
         async with async_session() as s2:  # 锁内重读 + 写入走新会话
             row = (await s2.execute(
                 select(StockCompany).where(StockCompany.ts_code == ts_code))).scalar_one_or_none()
-            if row is not None and _fresh(row.updated_at):
+            if row is not None and fresh(row.updated_at, _REFRESH_DAYS):
                 return _row_dict(row) if row.com_name is not None else None
             try:
                 fetched = await asyncio.to_thread(fetch_company_sync, ts_code)
@@ -174,5 +123,5 @@ async def get_company(session: AsyncSession, ts_code: str) -> dict | None:
                 if row is not None and row.com_name is not None:
                     return _row_dict(row)
                 raise
-            await _upsert(s2, fetched or {"ts_code": ts_code, "updated_at": _utcnow()})
+            await _upsert(s2, fetched or {"ts_code": ts_code, "updated_at": utcnow()})
             return fetched