fix(HEL-494): 数据中枢独占调度,主网站不再回退旧接口

主网站只向中枢要业务数据;来源选择、切源、补数全部在中枢内部完成,失败不再走东财/腾讯/Tushare 保底。

Co-authored-by: Cursor <cursoragent@cursor.com>
Co-authored-by: multica-agent <github@multica.ai>
This commit is contained in:
总工
2026-09-08 21:43:31 +08:00
co-authored by Cursor multica-agent
parent ef13d6feb5
commit 0b8419abca
23 changed files with 1159 additions and 368 deletions
+80 -47
View File
@@ -19,6 +19,7 @@ from backend.data.datahub.redact import redact_text, redact_value
from backend.data.datahub.route_state import LEDGER
from backend.data.datahub.settings import DatahubSettings
from backend.data.providers.tushare_client import TushareClient
from backend.data.providers.tushare_transport import TushareError
LOGGER = logging.getLogger("xiaobai.datahub")
ShadowSink = Callable[[dict[str, Any]], None]
@@ -171,6 +172,45 @@ class DatahubBridge:
self._log_failure("index_quotes", exc)
return None
def try_sector_quote(self, code: str, trade_date: str = "") -> dict[str, Any] | None:
flags = self.settings.flags("quotes")
if not flags.read:
return None
try:
response = self.client.sector_quote(code, trade_date)
data = response.data
if not isinstance(data, dict) or not data:
raise DatahubError("EMPTY", "datahub sector quote empty")
row = dict(data)
if (response.meta or {}).get("stale"):
row["delayed"] = True
row["delay_seconds"] = int((response.meta or {}).get("staleness_seconds") or 0)
row["delay_notice"] = str((response.meta or {}).get("delay_notice") or "")
self._record_route("quotes", "datahub", str((response.meta or {}).get("source") or "datahub"))
return row
except Exception as exc:
self._log_failure("quotes", exc)
return None
def try_limit_pool(self, trade_date: str = "") -> list[dict[str, Any]] | None:
flags = self.settings.flags("limit_events")
if not flags.read:
return None
try:
response = self.client.limit_pool(trade_date)
rows = [dict(item) for item in (response.data or []) if isinstance(item, dict)]
if not rows:
raise DatahubError("EMPTY", "datahub limit pool empty")
self._record_route(
"limit_events",
"datahub",
str((response.meta or {}).get("source") or "datahub"),
)
return rows
except Exception as exc:
self._log_failure("limit_events", exc)
return None
def try_daily_chart(
self,
code: str,
@@ -191,6 +231,11 @@ class DatahubBridge:
self.client.index_bars,
{"code": code, "from": start, "to": compact_end},
)
elif dataset == "sector_daily":
response = self._paginate(
self.client.sectors,
{"code": code, "from": start, "to": compact_end},
)
else:
response = self._paginate(
self.client.daily_bars,
@@ -263,53 +308,33 @@ class DatahubBridge:
fields: str,
legacy_query: Callable[..., list[dict[str, Any]]],
) -> list[dict[str, Any]]:
del legacy_query # 主网站不再直连 Tushare;调度全部由数据中枢完成。
if api_name == "rt_sw_k":
raise TushareError("rt_sw_k is disabled; use published sw_daily or free Shenwan realtime")
dataset = API_TO_DATASET.get(api_name)
# 问天按实际数据依赖接入:已映射到 hub 的 API 跟随开关;未映射的继续旧链路。
if not dataset:
return legacy_query(api_name, params, fields)
flags = self.settings.flags(dataset)
if not flags.read and not flags.shadow:
return legacy_query(api_name, params, fields)
hub_rows: list[dict[str, Any]] | None = None
hub_meta: dict[str, Any] = {}
hub_error: str | None = None
hub_canonical: list[dict[str, Any]] = []
try:
response = self._fetch_dataset(dataset, params or {}, api_name=api_name)
hub_canonical = self._extract_rows(dataset, response, params or {})
hub_rows = to_native_rows(dataset, hub_canonical)
hub_meta = dict(response.meta)
self._validate_usable(dataset, hub_rows, response)
except Exception as exc:
hub_error = self._error_text(exc)
self._log_failure(dataset, exc)
if flags.shadow:
try:
legacy_rows = legacy_query(api_name, params, fields)
except Exception as exc:
if flags.read and hub_rows is not None and hub_error is None:
self._emit_shadow(
compare_rows(dataset, [], hub_canonical, hub_meta, self._error_text(exc), fields)
)
return project_fields(hub_rows, fields)
raise
self._emit_shadow(compare_rows(dataset, legacy_rows, hub_canonical, hub_meta, hub_error, fields))
if flags.read and hub_rows is not None and hub_error is None:
self._record_route(dataset, "datahub", str(hub_meta.get("source") or "datahub"))
return project_fields(hub_rows, fields)
if dataset:
flags = self.settings.flags(dataset)
if flags.read:
self._record_route(dataset, "legacy", "tushare", hub_error or "")
return legacy_rows
if flags.read and hub_rows is not None and hub_error is None:
self._record_route(dataset, "datahub", str(hub_meta.get("source") or "datahub"))
return project_fields(hub_rows, fields)
result = legacy_query(api_name, params, fields)
if flags.read:
self._record_route(dataset, "legacy", "tushare", hub_error or "")
return result
try:
response = self._fetch_dataset(dataset, params or {}, api_name=api_name)
hub_canonical = self._extract_rows(dataset, response, params or {})
hub_rows = to_native_rows(dataset, hub_canonical)
self._validate_usable(dataset, hub_rows, response)
self._record_route(dataset, "datahub", str(response.meta.get("source") or "datahub"))
return project_fields(hub_rows, fields)
except Exception as exc:
self._log_failure(dataset, exc)
try:
response = self.client.query_api(api_name, params or {}, fields)
rows = [dict(item) for item in (response.data or []) if isinstance(item, dict)]
if dataset:
self._record_route(dataset, "datahub", str((response.meta or {}).get("source") or "datahub"))
else:
self._record_route(api_name, "datahub", str((response.meta or {}).get("source") or "datahub"))
return rows if not fields else project_fields(rows, fields)
except Exception as exc:
self._log_failure(dataset or api_name, exc)
raise TushareError(self._error_text(exc)) from exc
def _fetch_dataset(self, dataset: str, params: dict[str, Any], api_name: str = "") -> DatahubResponse:
date = yyyymmdd(params.get("trade_date") or params.get("date"))
@@ -429,8 +454,8 @@ class DatahubBridge:
def _log_failure(self, dataset: str, exc: Exception) -> None:
error = redact_text(self._error_text(exc), self.settings.secrets())
LOGGER.warning("datahub fallback dataset=%s error=%s", dataset, error)
self._record_route(dataset, "legacy", "pending-legacy", error)
LOGGER.warning("datahub unavailable dataset=%s error=%s", dataset, error)
self._record_route(dataset, "datahub", "unavailable", error)
def _record_route(self, dataset: str, route: str, source: str = "", error: str = "") -> None:
LEDGER.record(dataset, route, source, redact_text(error, self.settings.secrets()))
@@ -529,6 +554,8 @@ class DatahubAwareTushareClient:
legacy.try_market_quotes = self.try_market_quotes
legacy.try_quotes = self.try_quotes
legacy.try_index_quotes = self.try_index_quotes
legacy.try_sector_quote = self.try_sector_quote
legacy.try_limit_pool = self.try_limit_pool
legacy.record_datahub_legacy = self.record_datahub_legacy
def query(
@@ -548,6 +575,12 @@ class DatahubAwareTushareClient:
def try_index_quotes(self) -> list[dict[str, Any]] | None:
return self._bridge.try_index_quotes()
def try_sector_quote(self, code: str, trade_date: str = "") -> dict[str, Any] | None:
return self._bridge.try_sector_quote(code, trade_date)
def try_limit_pool(self, trade_date: str = "") -> list[dict[str, Any]] | None:
return self._bridge.try_limit_pool(trade_date)
def record_datahub_legacy(self, dataset: str, source: str = "", error: str = "") -> None:
self._bridge.record_legacy(dataset, source, error)