fix: correct ETF secid routing for efinance (#1535)

Co-authored-by: mumu <42829555+ZhuLinsen@users.noreply.github.com>

data_provider/efinance_fetcher.py

@@ -53,7 +53,18 @@ except (ValueError, TypeError):
 
 from src.patches.eastmoney_patch import eastmoney_patch
 from src.config import get_config
-from .base import BaseFetcher, DataFetchError, RateLimitError, STANDARD_COLUMNS,is_bse_code, is_st_stock, is_kc_cy_stock, normalize_stock_code, _is_hk_market
+from .base import (
+    BaseFetcher,
+    DataFetchError,
+    RateLimitError,
+    STANDARD_COLUMNS,
+    is_bse_code,
+    is_st_stock,
+    is_kc_cy_stock,
+    normalize_stock_code,
+    _is_hk_market,
+    _is_etf_code as _is_a_share_etf_code,
+)
 from .realtime_types import (
     UnifiedRealtimeQuote, RealtimeSource,
     get_realtime_circuit_breaker,
@@ -134,6 +145,9 @@ _etf_realtime_cache: Dict[str, Any] = {
     'ttl': 600  # 10分钟缓存有效期
 }
 
+_ETF_SH_PREFIXES = ('51', '52', '56', '58')
+_ETF_SZ_PREFIXES = ('15', '16', '18')
+
 
 def _is_etf_code(stock_code: str) -> bool:
     """
@@ -149,8 +163,19 @@ def _is_etf_code(stock_code: str) -> bool:
     Returns:
         True 表示是 ETF 代码，False 表示是普通股票代码
     """
-    etf_prefixes = ('51', '52', '56', '58', '15', '16', '18')
-    return stock_code.startswith(etf_prefixes) and len(stock_code) == 6
+    return _is_a_share_etf_code(stock_code)
+
+
+def _build_eastmoney_etf_secid(stock_code: str) -> str:
+    """Build Eastmoney secid for A-share ETF historical K-line queries."""
+    code = normalize_stock_code(stock_code)
+    if not _is_etf_code(code):
+        raise DataFetchError(f"无法识别 ETF 代码 {stock_code}")
+    if code.startswith(_ETF_SH_PREFIXES):
+        return f"1.{code}"
+    if code.startswith(_ETF_SZ_PREFIXES):
+        return f"0.{code}"
+    raise DataFetchError(f"无法确定 ETF {stock_code} 的 Eastmoney 市场前缀")
 
 
 def _is_us_code(stock_code: str) -> bool:
@@ -480,20 +505,26 @@ class EfinanceFetcher(BaseFetcher):
         # Format dates (efinance uses YYYYMMDD)
         beg_date = start_date.replace('-', '')
         end_date_fmt = end_date.replace('-', '')
+        secid = _build_eastmoney_etf_secid(stock_code)
 
-        logger.info(f"[API调用] ef.stock.get_quote_history(stock_codes={stock_code}, "
-                     f"beg={beg_date}, end={end_date_fmt}, klt=101, fqt=1)  [ETF]")
+        logger.info(
+            f"[API调用] ef.stock.get_quote_history(stock_codes={secid}, "
+            f"beg={beg_date}, end={end_date_fmt}, klt=101, fqt=1, "
+            f"quote_id_mode=True, use_id_cache=False)  [ETF stock_code={stock_code}]"
+        )
 
         api_start = time.time()
         try:
             # ETFs are exchange-traded securities; use the stock API to get full OHLCV data
             df = _ef_call_with_timeout(
                 ef.stock.get_quote_history,
-                stock_codes=stock_code,
+                stock_codes=secid,
                 beg=beg_date,
                 end=end_date_fmt,
                 klt=101,  # daily
                 fqt=1,    # forward-adjusted
+                quote_id_mode=True,
+                use_id_cache=False,
                 timeout=60,
             )
 
@@ -502,7 +533,7 @@ class EfinanceFetcher(BaseFetcher):
             if df is not None and not df.empty:
                 logger.info(
                     "[API返回] Eastmoney 历史K线成功 [ETF]: "
-                    f"endpoint={EASTMONEY_HISTORY_ENDPOINT}, stock_code={stock_code}, "
+                    f"endpoint={EASTMONEY_HISTORY_ENDPOINT}, stock_code={stock_code}, secid={secid}, "
                     f"range={beg_date}~{end_date_fmt}, rows={len(df)}, elapsed={api_elapsed:.2f}s"
                 )
                 logger.info(f"[API返回] 列名: {list(df.columns)}")
@@ -512,7 +543,7 @@ class EfinanceFetcher(BaseFetcher):
             else:
                 logger.warning(
                     "[API返回] Eastmoney 历史K线为空 [ETF]: "
-                    f"endpoint={EASTMONEY_HISTORY_ENDPOINT}, stock_code={stock_code}, "
+                    f"endpoint={EASTMONEY_HISTORY_ENDPOINT}, stock_code={stock_code}, secid={secid}, "
                     f"range={beg_date}~{end_date_fmt}, elapsed={api_elapsed:.2f}s"
                 )
 

docs/CHANGELOG.md

@@ -22,6 +22,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/).
 - [修复] Windows 桌面端自动更新传给 NSIS 的 `/D=` 目录参数在包含空格时自动加引号，避免安装位置注册表被截断。
 
 - [修复] 加固 LLM channel base_url 校验，避免解析差异导致 SSRF 绕过。
+- [修复] 修正 efinance ETF 日线 Eastmoney secid 路由，避免沪市 ETF 被按深市 quote id 查询导致日线为空。
+- [测试] 补充 ETF 日线数据源路由、输入变体、fallback 与 MA 字段回归覆盖。
 - [改进] 优化 Web 报告详情页信息层级，将输入数据块和运行诊断下移为主体内容后的折叠辅助信息。
 - [新功能] 市场阶段低敏摘要接入历史详情、同步分析响应和 completed 任务状态的 report metadata。
 

tests/test_etf_daily_routing.py

@@ -0,0 +1,203 @@
+import os
+import sys
+import types
+from types import SimpleNamespace
+from unittest.mock import MagicMock, patch
+
+import pandas as pd
+
+sys.path.insert(0, os.path.abspath(os.path.join(os.path.dirname(__file__), "..")))
+
+from data_provider.akshare_fetcher import AkshareFetcher
+from data_provider.base import BaseFetcher, DataFetchError, DataFetcherManager
+from data_provider.efinance_fetcher import EfinanceFetcher
+
+
+def _make_efinance_fetcher() -> EfinanceFetcher:
+    with patch(
+        "data_provider.efinance_fetcher.get_config",
+        return_value=SimpleNamespace(enable_eastmoney_patch=False),
+    ):
+        return EfinanceFetcher(sleep_min=0, sleep_max=0)
+
+
+def _make_akshare_fetcher() -> AkshareFetcher:
+    with patch(
+        "data_provider.akshare_fetcher.get_config",
+        return_value=SimpleNamespace(enable_eastmoney_patch=False),
+    ):
+        return AkshareFetcher(sleep_min=0, sleep_max=0)
+
+
+def _history_frame(code: str = "563230") -> pd.DataFrame:
+    return pd.DataFrame(
+        {
+            "股票代码": [code] * 5,
+            "日期": pd.date_range("2026-01-01", periods=5).strftime("%Y-%m-%d"),
+            "开盘": [10.0, 10.1, 10.2, 10.3, 10.4],
+            "收盘": [10.1, 10.2, 10.3, 10.4, 10.5],
+            "最高": [10.2, 10.3, 10.4, 10.5, 10.6],
+            "最低": [9.9, 10.0, 10.1, 10.2, 10.3],
+            "成交量": [1000, 1100, 1200, 1300, 1400],
+            "成交额": [10100, 11220, 12360, 13520, 14700],
+            "涨跌幅": [0.0, 0.99, 0.98, 0.97, 0.96],
+        }
+    )
+
+
+def _run_efinance_daily(stock_code: str) -> tuple[pd.DataFrame, MagicMock]:
+    fetcher = _make_efinance_fetcher()
+    fake_efinance = types.SimpleNamespace(
+        stock=types.SimpleNamespace(get_quote_history=MagicMock(name="get_quote_history"))
+    )
+    call = MagicMock(return_value=_history_frame())
+
+    with patch.dict(sys.modules, {"efinance": fake_efinance}):
+        with patch("data_provider.efinance_fetcher._ef_call_with_timeout", call):
+            with patch.object(fetcher, "_set_random_user_agent"), patch.object(
+                fetcher, "_enforce_rate_limit"
+            ):
+                df = fetcher.get_daily_data(
+                    stock_code,
+                    start_date="2026-01-01",
+                    end_date="2026-01-05",
+                )
+
+    return df, call
+
+
+def test_efinance_sh_etf_uses_eastmoney_quote_id_mode() -> None:
+    df, call = _run_efinance_daily("563230")
+
+    kwargs = call.call_args.kwargs
+    assert kwargs["stock_codes"] == "1.563230"
+    assert kwargs["quote_id_mode"] is True
+    assert kwargs["use_id_cache"] is False
+    assert kwargs["beg"] == "20260101"
+    assert kwargs["end"] == "20260105"
+    assert kwargs["klt"] == 101
+    assert kwargs["fqt"] == 1
+    assert {"ma5", "ma10", "ma20"}.issubset(df.columns)
+
+
+def test_efinance_sz_etf_uses_eastmoney_quote_id_mode() -> None:
+    _, call = _run_efinance_daily("159919")
+
+    assert call.call_args.kwargs["stock_codes"] == "0.159919"
+    assert call.call_args.kwargs["quote_id_mode"] is True
+    assert call.call_args.kwargs["use_id_cache"] is False
+
+
+def test_efinance_etf_code_variants_use_sh_secid() -> None:
+    for stock_code in ("SH563230", "SH.563230", "563230.SH"):
+        _, call = _run_efinance_daily(stock_code)
+        assert call.call_args.kwargs["stock_codes"] == "1.563230"
+
+
+def test_akshare_etf_uses_fund_etf_hist_em() -> None:
+    fetcher = _make_akshare_fetcher()
+    fake_akshare = types.SimpleNamespace(fund_etf_hist_em=MagicMock(return_value=_history_frame()))
+
+    with patch.dict(sys.modules, {"akshare": fake_akshare}):
+        with patch.object(fetcher, "_set_random_user_agent"), patch.object(
+            fetcher, "_enforce_rate_limit"
+        ):
+            df = fetcher._fetch_raw_data("563230", "2026-01-01", "2026-01-05")
+
+    assert df is not None
+    fake_akshare.fund_etf_hist_em.assert_called_once_with(
+        symbol="563230",
+        period="daily",
+        start_date="20260101",
+        end_date="20260105",
+        adjust="qfq",
+    )
+
+
+def test_manager_normalizes_prefixed_etf_before_efinance_secid_route() -> None:
+    fetcher = _make_efinance_fetcher()
+    manager = DataFetcherManager(fetchers=[fetcher])
+    fake_efinance = types.SimpleNamespace(
+        stock=types.SimpleNamespace(get_quote_history=MagicMock(name="get_quote_history"))
+    )
+    call = MagicMock(return_value=_history_frame())
+
+    with patch.dict(sys.modules, {"efinance": fake_efinance}):
+        with patch("data_provider.efinance_fetcher._ef_call_with_timeout", call):
+            with patch.object(fetcher, "_set_random_user_agent"), patch.object(
+                fetcher, "_enforce_rate_limit"
+            ):
+                df, source = manager.get_daily_data(
+                    "SH563230",
+                    start_date="2026-01-01",
+                    end_date="2026-01-05",
+                )
+
+    assert source == "EfinanceFetcher"
+    assert call.call_args.kwargs["stock_codes"] == "1.563230"
+    assert {"ma5", "ma10", "ma20"}.issubset(df.columns)
+
+
+class _EmptyEfinanceFetcher(BaseFetcher):
+    name = "EfinanceFetcher"
+    priority = 0
+
+    def _fetch_raw_data(self, stock_code: str, start_date: str, end_date: str) -> pd.DataFrame:
+        return pd.DataFrame()
+
+    def _normalize_data(self, df: pd.DataFrame, stock_code: str) -> pd.DataFrame:
+        return df
+
+
+class _FailingEfinanceFetcher(BaseFetcher):
+    name = "EfinanceFetcher"
+    priority = 0
+
+    def _fetch_raw_data(self, stock_code: str, start_date: str, end_date: str) -> pd.DataFrame:
+        raise DataFetchError("efinance ETF history failed")
+
+    def _normalize_data(self, df: pd.DataFrame, stock_code: str) -> pd.DataFrame:
+        return df
+
+
+class _SuccessfulAkshareFetcher(BaseFetcher):
+    name = "AkshareFetcher"
+    priority = 1
+
+    def _fetch_raw_data(self, stock_code: str, start_date: str, end_date: str) -> pd.DataFrame:
+        return pd.DataFrame(
+            {
+                "code": [stock_code] * 5,
+                "date": pd.date_range("2026-01-01", periods=5),
+                "open": [10.0, 10.1, 10.2, 10.3, 10.4],
+                "high": [10.2, 10.3, 10.4, 10.5, 10.6],
+                "low": [9.9, 10.0, 10.1, 10.2, 10.3],
+                "close": [10.1, 10.2, 10.3, 10.4, 10.5],
+                "volume": [1000, 1100, 1200, 1300, 1400],
+                "amount": [10100, 11220, 12360, 13520, 14700],
+                "pct_chg": [0.0, 0.99, 0.98, 0.97, 0.96],
+            }
+        )
+
+    def _normalize_data(self, df: pd.DataFrame, stock_code: str) -> pd.DataFrame:
+        return df
+
+
+def test_manager_falls_back_and_keeps_etf_ma_columns() -> None:
+    manager = DataFetcherManager(fetchers=[_EmptyEfinanceFetcher(), _SuccessfulAkshareFetcher()])
+
+    df, source = manager.get_daily_data("563230", start_date="2026-01-01", end_date="2026-01-05")
+
+    assert source == "AkshareFetcher"
+    assert {"ma5", "ma10", "ma20"}.issubset(df.columns)
+    assert df["ma5"].iloc[-1] == 10.3
+
+
+def test_manager_falls_back_when_efinance_raises_and_keeps_etf_ma_columns() -> None:
+    manager = DataFetcherManager(fetchers=[_FailingEfinanceFetcher(), _SuccessfulAkshareFetcher()])
+
+    df, source = manager.get_daily_data("563230", start_date="2026-01-01", end_date="2026-01-05")
+
+    assert source == "AkshareFetcher"
+    assert {"ma5", "ma10", "ma20"}.issubset(df.columns)
+    assert df["ma5"].iloc[-1] == 10.3

tests/test_tushare_fetcher_get_stock_list.py

@@ -142,6 +142,22 @@ class TestTushareFetcherFetchRawData(unittest.TestCase):
         fetcher._api.daily.assert_not_called()
         fetcher._api.hk_daily.assert_not_called()
 
+    def test_fetch_raw_data_56_prefix_etf_uses_sh_fund_daily(self) -> None:
+        fetcher = self._make_fetcher()
+        fetcher._api.fund_daily.return_value = pd.DataFrame({"trade_date": ["20260101"]})
+
+        with patch.object(fetcher, "_check_rate_limit"):
+            out = fetcher._fetch_raw_data("563230", "20260101", "20260105")
+
+        self.assertIsNotNone(out)
+        fetcher._api.fund_daily.assert_called_once_with(
+            ts_code="563230.SH",
+            start_date="20260101",
+            end_date="20260105",
+        )
+        fetcher._api.daily.assert_not_called()
+        fetcher._api.hk_daily.assert_not_called()
+
     def test_fetch_raw_data_hk_uses_hk_daily(self) -> None:
         fetcher = self._make_fetcher()
         fetcher._api.hk_daily.return_value = pd.DataFrame({"trade_date": ["20260102"]})
@@ -274,4 +290,3 @@ class TestTushareFetcherChipDistribution(unittest.TestCase):
 
 if __name__ == "__main__":
     unittest.main()
-