fix(HEL-190): 按真实交易日历补齐最近60日快照,修复断档后只显示当天

保留连续性过滤,新增可审计补档工具与备份步骤;周末/节假日与真缺档分开处理,支持重复执行与部分失败续跑。

Co-authored-by: Cursor <cursoragent@cursor.com>
Co-authored-by: multica-agent <github@multica.ai>
This commit is contained in:
施工员
2026-08-27 14:50:32 +00:00
co-authored by Cursor multica-agent
parent 8a5e78f022
commit 7ad445bc9f
10 changed files with 974 additions and 31 deletions
+229 -21
View File
@@ -3,9 +3,11 @@ from __future__ import annotations
import copy
import re
from datetime import date, datetime, time as dt_time, timedelta
from pathlib import Path
from typing import Any
from backend.bootstrap.config import (
DATA_DIR,
normalize_date,
tushare_code,
validate_stock_code,
@@ -13,6 +15,17 @@ from backend.bootstrap.config import (
)
from backend.data.providers.ifind_client import IfindError
from backend.data.providers.tushare_client import TushareClient, TushareError
from backend.features.market.backfill_history import (
DEFAULT_RECENT_TRADING_DAYS,
MAX_RANGE_TRADING_DAYS,
build_backfill_audit,
calendar_window_start,
classify_snapshot_coverage,
create_sqlite_backup,
display_date,
select_open_trade_dates,
select_open_trade_dates_in_range,
)
from backend.features.market.charts import ChartDataError
from backend.features.market.insights import MarketInsightsService
from backend.features.sentiment.engine import SENTIMENT_ENGINE_VERSION
@@ -890,31 +903,226 @@ class MarketServiceMixin:
"intraday": intraday_points,
}
def backfill(self, start_date: str, end_date: str) -> list[dict[str, Any]]:
start = datetime.strptime(normalize_date(start_date), "%Y%m%d").date()
end = datetime.strptime(normalize_date(end_date), "%Y%m%d").date()
if start > end:
raise ValueError("开始日期不能晚于结束日期。")
weekdays = []
current = start
while current <= end:
if current.weekday() < 5:
weekdays.append(current)
current += timedelta(days=1)
if len(weekdays) > 15:
raise ValueError("单次最多回补 15 个工作日。")
results = []
for day in weekdays:
dashboard = self.sync_dashboard(day.strftime("%Y%m%d"))
def backfill(
self,
start_date: str = "",
end_date: str = "",
*,
lookback: int | None = None,
dry_run: bool = False,
force: bool = False,
create_backup: bool = True,
) -> dict[str, Any]:
"""Backfill dashboard snapshots for real trading days only.
- Date-range mode keeps the admin UI contract (max 15 open sessions).
- Recent mode fills the last N open sessions (default/max 60).
Weekends and holidays are reported as skipped non-trading days, not errors.
"""
if not self.configured:
raise ValueError("公共行情尚未配置,无法回补历史快照。")
normalized_end = normalize_date(end_date or date.today().isoformat())
if lookback is not None or not (start_date and end_date):
target_lookback = (
DEFAULT_RECENT_TRADING_DAYS if lookback is None else int(lookback)
)
return self.backfill_recent_trading_days(
end_date=normalized_end,
lookback=target_lookback,
dry_run=dry_run,
force=force,
create_backup=create_backup,
)
return self._backfill_date_range(
start_date=normalize_date(start_date),
end_date=normalized_end,
dry_run=dry_run,
force=force,
create_backup=create_backup,
)
def backfill_recent_trading_days(
self,
end_date: str = "",
lookback: int = DEFAULT_RECENT_TRADING_DAYS,
*,
dry_run: bool = False,
force: bool = False,
create_backup: bool = True,
) -> dict[str, Any]:
normalized_end = normalize_date(end_date or date.today().isoformat())
trade_dates = self._load_recent_open_trade_dates(normalized_end, lookback)
existing = self.database.list_snapshot_trade_dates(
trade_dates[0], trade_dates[-1]
)
coverage = classify_snapshot_coverage(trade_dates, existing)
return self._execute_snapshot_backfill(
mode="recent",
end_date=normalized_end,
lookback=lookback,
coverage=coverage,
skipped_non_trading_days=[],
dry_run=dry_run,
force=force,
create_backup=create_backup,
)
def _backfill_date_range(
self,
start_date: str,
end_date: str,
*,
dry_run: bool = False,
force: bool = False,
create_backup: bool = True,
) -> dict[str, Any]:
window_start = calendar_window_start(end_date, MAX_RANGE_TRADING_DAYS)
calendar_rows = self._tushare_client().query(
"trade_cal",
{
"exchange": "SSE",
"start_date": min(window_start, start_date),
"end_date": end_date,
},
"cal_date,is_open,pretrade_date",
)
trade_dates, skipped = select_open_trade_dates_in_range(
calendar_rows,
start_date,
end_date,
maximum=MAX_RANGE_TRADING_DAYS,
)
if not trade_dates:
raise ValueError("选定区间内没有交易日,周末或节假日无需回补。")
existing = self.database.list_snapshot_trade_dates(trade_dates[0], trade_dates[-1])
coverage = classify_snapshot_coverage(trade_dates, existing)
return self._execute_snapshot_backfill(
mode="range",
end_date=end_date,
lookback=None,
coverage=coverage,
skipped_non_trading_days=skipped,
dry_run=dry_run,
force=force,
create_backup=create_backup,
)
def _load_recent_open_trade_dates(self, end_date: str, lookback: int) -> list[str]:
start_date = calendar_window_start(end_date, lookback)
calendar_rows = self._tushare_client().query(
"trade_cal",
{
"exchange": "SSE",
"start_date": start_date,
"end_date": end_date,
},
"cal_date,is_open,pretrade_date",
)
return select_open_trade_dates(calendar_rows, end_date, lookback)
def _execute_snapshot_backfill(
self,
*,
mode: str,
end_date: str,
lookback: int | None,
coverage: dict[str, Any],
skipped_non_trading_days: list[str],
dry_run: bool,
force: bool,
create_backup: bool,
) -> dict[str, Any]:
targets = list(coverage["trade_dates"] if force else coverage["missing"])
backup_path: str | None = None
if create_backup and not dry_run and targets:
backup = create_sqlite_backup(
Path(self.database.path),
DATA_DIR / "backups",
label=f"pre-{mode}-backfill",
)
backup_path = str(backup)
results: list[dict[str, Any]] = []
if dry_run:
for trade_date in coverage["trade_dates"]:
exists = trade_date in coverage["present"]
if exists and not force:
status = "skipped"
action = "exists"
else:
status = "planned"
action = "refresh" if exists else "create"
results.append(
{
"requested_date": display_date(trade_date),
"trade_date": display_date(trade_date),
"status": status,
"action": action,
}
)
return build_backfill_audit(
mode=mode,
end_date=end_date,
lookback=lookback,
coverage=coverage,
skipped_non_trading_days=skipped_non_trading_days,
backup_path=backup_path,
dry_run=True,
results=results,
)
present_before = set(coverage["present"])
for trade_date in targets:
existed = trade_date in present_before
try:
dashboard = self.sync_dashboard(trade_date)
actual = normalize_date(
str(dashboard.get("meta", {}).get("trade_date") or trade_date)
)
results.append(
{
"requested_date": display_date(trade_date),
"trade_date": display_date(actual),
"status": "success",
"action": "refreshed" if existed else "created",
"source": dashboard.get("meta", {}).get("source"),
"records": self._record_count(dashboard),
}
)
except Exception as exc:
results.append(
{
"requested_date": display_date(trade_date),
"trade_date": display_date(trade_date),
"status": "failed",
"action": "refresh" if existed else "create",
"error": str(exc),
}
)
for trade_date in coverage["present"]:
if force:
continue
results.append(
{
"requested_date": day.isoformat(),
"trade_date": dashboard["meta"]["trade_date"],
"source": dashboard["meta"]["source"],
"records": self._record_count(dashboard),
"requested_date": display_date(trade_date),
"trade_date": display_date(trade_date),
"status": "skipped",
"action": "exists",
}
)
return results
results.sort(key=lambda row: str(row.get("requested_date") or ""))
return build_backfill_audit(
mode=mode,
end_date=end_date,
lookback=lookback,
coverage=coverage,
skipped_non_trading_days=skipped_non_trading_days,
backup_path=backup_path,
dry_run=False,
results=results,
)
def _stock_identity(self, code: str, trade_date: str) -> tuple[str, str]:
snapshot = self.database.get_snapshot(trade_date) or {}