提交

backend/app/data/sync_utils.py

@@ -0,0 +1,87 @@
+"""懒加载管道共用件：tushare pro 单例、限频重试、取值清洗、stock_sync_state 读写。
+
+finance / dividend 等按需单查管道共用；语义与 company.py 内的私有版本一致
+（那处历史代码未迁移，新管道一律从这里取）。
+"""
+from __future__ import annotations
+
+import time
+from datetime import datetime, timedelta, timezone
+
+from sqlalchemy import select
+from sqlalchemy.dialects.postgresql import insert as pg_insert
+from sqlalchemy.ext.asyncio import AsyncSession
+
+from ..config import settings
+from ..models import StockSyncState
+
+# 频率超限特征（等待 62s 重试一次；与 company.py / data.etf_sync._call_retry 同款语义）
+_RATE_MARKS = ("频率超限", "每分钟")
+
+_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_lazy():
+    if not settings.tushare_token:
+        raise RuntimeError("未配置 TUSHARE_TOKEN，无法拉取 tushare 数据（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, days: int) -> bool:
+    return updated_at is not None and updated_at >= utcnow() - timedelta(days=days)
+
+
+def s_clean(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_clean(v) -> float | None:
+    try:
+        f = float(v)
+    except (TypeError, ValueError):
+        return None
+    return None if f != f else f  # NaN -> None
+
+
+async def read_sync_state(session: AsyncSession, ts_code: str, kind: str) -> StockSyncState | None:
+    return (await session.execute(
+        select(StockSyncState).where(
+            StockSyncState.ts_code == ts_code, StockSyncState.kind == kind
+        ))).scalar_one_or_none()
+
+
+async def upsert_sync_state(session: AsyncSession, ts_code: str, kind: str, *, has_data: bool) -> None:
+    """写入同步状态（不 commit，由调用方统一提交）。"""
+    stmt = pg_insert(StockSyncState).values(
+        ts_code=ts_code, kind=kind, last_synced_at=utcnow(), has_data=has_data,
+    )
+    stmt = stmt.on_conflict_do_update(
+        index_elements=["ts_code", "kind"],
+        set_={"last_synced_at": stmt.excluded.last_synced_at, "has_data": stmt.excluded.has_data},
+    )
+    await session.execute(stmt)