qlib/backend/app/infrastructure/data_sources/failover.py

"""Tushare → Sina Failover 包装（AGENT.md §7）。

规则：
- 优先 primary；primary 抛错时才尝试 fallback（避免对空结果做无谓兜底请求）
- fallback 不支持该 API（DataSourceNotSupported）或自身失败 → 抛 DataSourceError
- 每次尝试都写 SyncLog（source / 成功与否 / 行数 / 区间），禁止静默切换
"""

from __future__ import annotations

from collections.abc import Callable
from datetime import date
from typing import Any

from app.domain.entities.market import SyncLog
from app.domain.providers import MarketDataProvider
from app.infrastructure.data_sources.errors import DataSourceError, DataSourceNotSupported


class FailoverProvider:
    """以 primary 为主、fallback 为辅的 MarketDataProvider 实现。"""

    name = "failover"

    def __init__(
        self,
        primary: MarketDataProvider,
        fallback: MarketDataProvider | None = None,
        *,
        audit: Callable[[SyncLog], None] | None = None,
    ) -> None:
        self.primary = primary
        self.fallback = fallback
        self._audit = audit or (lambda _log: None)

    # ---- 各 API 代理 ----

    def get_stock_basic(self, list_status: str = "L") -> list:
        return self._with_failover(
            "get_stock_basic",
            primary_call=lambda: self.primary.get_stock_basic(list_status),
            fallback_call=lambda: self.fallback.get_stock_basic(list_status),
        )

    def get_name_changes(self, start: date, end: date) -> list:
        return self._with_failover(
            "get_name_changes",
            start=start,
            end=end,
            primary_call=lambda: self.primary.get_name_changes(start, end),
            fallback_call=lambda: self.fallback.get_name_changes(start, end),
        )

    def get_trade_cal(self, start: date, end: date) -> list:
        return self._with_failover(
            "get_trade_cal",
            start=start,
            end=end,
            primary_call=lambda: self.primary.get_trade_cal(start, end),
            fallback_call=lambda: self.fallback.get_trade_cal(start, end),
        )

    def get_daily(self, symbol: str, start: date, end: date) -> list:
        return self._with_failover(
            "get_daily",
            start=start,
            end=end,
            primary_call=lambda: self.primary.get_daily(symbol, start, end),
            fallback_call=lambda: self.fallback.get_daily(symbol, start, end),
        )

    def get_index_weight(self, index_code: str) -> list:
        return self._with_failover(
            "get_index_weight",
            primary_call=lambda: self.primary.get_index_weight(index_code),
            fallback_call=lambda: self.fallback.get_index_weight(index_code),
        )

    def get_adjust_factor(self, symbol: str, start: date, end: date) -> list:
        return self._with_failover(
            "get_adjust_factor",
            start=start,
            end=end,
            primary_call=lambda: self.primary.get_adjust_factor(symbol, start, end),
            fallback_call=lambda: self.fallback.get_adjust_factor(symbol, start, end),
        )

    def get_financial(
        self, symbol: str, start: date | None = None, end: date | None = None
    ) -> list:
        return self._with_failover(
            "get_financial",
            start=start,
            end=end,
            primary_call=lambda: self.primary.get_financial(symbol, start, end),
            fallback_call=lambda: self.fallback.get_financial(symbol, start, end),
        )

    def get_daily_basic(self, trade_date: date) -> list:
        """每日指标：新浪不支持本接口 → 主源失败时如实报错，不做假兜底。"""
        return self._with_failover(
            "get_daily_basic",
            start=trade_date,
            end=trade_date,
            primary_call=lambda: self.primary.get_daily_basic(trade_date),
            fallback_call=(
                (lambda: self.fallback.get_daily_basic(trade_date))
                if self.fallback is not None
                else None
            ),
        )

    # ---- 内部 ----

    def _with_failover(
        self,
        api: str,
        *,
        primary_call: Callable[[], list],
        fallback_call: Callable[[], list] | None = None,
        start: date | None = None,
        end: date | None = None,
    ) -> list:
        try:
            rows = primary_call()
        except Exception as exc:  # noqa: BLE001 —— 统一走审计
            self._log(
                source=self.primary.name,
                api=api,
                success=False,
                reason=str(exc),
                start=start,
                end=end,
            )
            return self._try_fallback(api, fallback_call, start=start, end=end, primary_error=exc)
        self._log(
            source=self.primary.name,
            api=api,
            success=True,
            row_count=_len(rows),
            start=start,
            end=end,
        )
        return rows

    def _try_fallback(self, api, fallback_call, *, start, end, primary_error):
        if fallback_call is None or self.fallback is None:
            raise DataSourceError(
                f"{self.primary.name}.{api} 失败且无备用源: {primary_error}"
            ) from primary_error
        try:
            rows = fallback_call()
        except DataSourceNotSupported as exc:
            self._log(
                source=self.fallback.name,
                api=api,
                success=False,
                reason=f"不支持: {exc}",
                start=start,
                end=end,
            )
            raise DataSourceError(
                f"{self.primary.name}.{api} 失败，备用源不支持: {primary_error}"
            ) from primary_error
        except Exception as exc:  # noqa: BLE001
            self._log(
                source=self.fallback.name,
                api=api,
                success=False,
                reason=str(exc),
                start=start,
                end=end,
            )
            raise DataSourceError(
                f"主备数据源均失败: primary[{self.primary.name}]={primary_error} "
                f"fallback[{self.fallback.name}]={exc}"
            ) from exc
        self._log(
            source=self.fallback.name,
            api=api,
            success=True,
            row_count=_len(rows),
            start=start,
            end=end,
        )
        return rows

    def _log(
        self,
        *,
        source: str,
        api: str,
        success: bool,
        reason: str | None = None,
        row_count: int = 0,
        start: date | None = None,
        end: date | None = None,
    ) -> None:
        self._audit(
            SyncLog(
                source=source,
                api=api,
                success=success,
                failure_reason=reason,
                row_count=row_count,
                data_start=start,
                data_end=end,
            )
        )


def _len(rows: Any) -> int:
    return len(rows) if rows is not None else 0