Add official DOL and EIA economic calendar feeds

Collect weekly claims and precise commercial petroleum inventory changes with official schedule exceptions, distinct reporting periods, raw evidence, and revisions.

Extend explicit history backfills without changing their default source scope. Report archive gaps and reject mismatched weeks instead of filling current values into history.

backend/src/alphapulse/calendar/worker.py

@@ -12,7 +12,7 @@ from parsel import Selector
 from sqlalchemy import select
 
 from alphapulse.calendar.catalog import SCHEDULE_URLS, UMICH_CSV, UMICH_HISTORY, UMICH_HOME, UMICH_INFO
-from alphapulse.calendar import europe
+from alphapulse.calendar import dol, eia, europe
 from alphapulse.calendar.parsers import EASTERN, MONTHS, bea_release_link, parse_fed_calendar, parse_ics, parse_values, release_date, bls_periods, release_metadata
 from alphapulse.calendar.service import apply_bls_periods, record_revision, next_poll, save_raw, save_result, save_schedule, save_series
 from alphapulse.calendar.umich import current_result, document_links, monthly_series, parse_schedule, pdf_text
@@ -57,12 +57,14 @@ class CalendarWorker:
         self.http = httpx.AsyncClient(timeout=8, follow_redirects=True, headers={"User-Agent": "AlphaPulse economic calendar"})
         self.bls = httpx.AsyncClient(timeout=8, follow_redirects=True, verify=ctx, headers={"User-Agent": "AlphaPulse economic calendar"})
         self.sources: dict[str, dict] = {}
+        self.eia_schedule_body: str | None = None
         self.fetch_slots = asyncio.Semaphore(2)
 
     async def fetch(self, source: str, url: str, *, query: dict | None = None) -> tuple[httpx.Response, int]:
         # 链接来自公开响应，限制到对应官方域名，避免外部内容扩大访问范围。
         allowed = {"bls": {"www.bls.gov"}, "bea": {"www.bea.gov", "apps.bea.gov"}, "fed": {"www.federalreserve.gov"},
-                   "umich": {"www.sca.isr.umich.edu", "data.sca.isr.umich.edu"}, **HOSTS}
+                   "umich": {"www.sca.isr.umich.edu", "data.sca.isr.umich.edu"}, "dol": {"www.dol.gov", "oui.doleta.gov"},
+                   "eia": {"www.eia.gov", "ir.eia.gov"}, **HOSTS}
         if urlparse(url).scheme != "https" or urlparse(url).hostname not in allowed[source]:
             raise ValueError("公告链接不是对应机构的官方 HTTPS 地址")
         async with self.fetch_slots:
@@ -84,6 +86,8 @@ class CalendarWorker:
         urls += [("bls", f"https://www.bls.gov/schedule/news_release/{code}.htm") for code in ("cpi", "ppi", "empsit")]
         urls += [("umich", UMICH_INFO), ("umich", UMICH_CSV)]
         urls += list(europe.SCHEDULES.items())
+        urls += [("dol", dol.SCHEDULE)]
+        urls += [("eia", eia.SCHEDULE)]
         # 启动时明确各来源尚未就绪；月份滚动后移除不再检查的旧月状态。
         self.sources = {url: self.sources.get(url, {"source": source, "checked_at": now.isoformat(), "error": "尚未同步"})
                         for source, url in urls}
@@ -95,7 +99,14 @@ class CalendarWorker:
                 fetch_url = europe.eurostat_calendar_url(now) if source == "eurostat" else url
                 response, raw_id = await self.fetch(source, fetch_url, query=europe.INSEE_QUERY if source == "insee" else None)
                 response.raise_for_status()
-                if source in europe.SCHEDULES:
+                if source in ("dol", "eia"):
+                    events = (dol if source == "dol" else eia).schedule(response.text, now)
+                    async with SessionLocal() as session:
+                        await save_schedule(session, source, url, events, raw_id, datetime.now(UTC), withdraw_missing=False)
+                        await session.commit()
+                    if source == "eia":
+                        self.eia_schedule_body = response.text
+                elif source in europe.SCHEDULES:
                     events = europe.parse_schedule(source, response.text)
                     async with SessionLocal() as session:
                         # 欧洲公开日程是滚动/分页窗口，缺失不等于撤回。
@@ -168,6 +179,30 @@ class CalendarWorker:
             await session.commit()
 
     async def result_page(self, row: EconomicRelease) -> tuple[httpx.Response, int] | None:
+        if row.source == "eia":
+            if self.eia_schedule_body is None:
+                schedule, _ = await self.fetch("eia", eia.SCHEDULE)
+                schedule.raise_for_status()
+                eia.exceptions(schedule.text)
+                self.eia_schedule_body = schedule.text
+            response, raw_id = await self.fetch("eia", row.release_url or eia.LATEST)
+            response.raise_for_status()
+            announcement = eia.report(response.text, self.eia_schedule_body)
+            if announcement.at < row.scheduled_at or (announcement.at > row.scheduled_at and row.first_received_at is not None):
+                return None
+            if announcement.at != row.scheduled_at or announcement.period != row.period:
+                raise ValueError("EIA 公告统计周或时间与所选日程不符，未填入其他期次的数据")
+            return response, raw_id
+        if row.source == "dol":
+            response, raw_id = await self.fetch("dol", row.release_url or dol.LATEST)
+            response.raise_for_status()
+            announcements = await asyncio.to_thread(dol.report, await asyncio.to_thread(pdf_text, response.content))
+            announcement = announcements[row.family]
+            if announcement.at < row.scheduled_at or (announcement.at > row.scheduled_at and row.first_received_at is not None):
+                return None
+            if announcement.at != row.scheduled_at:
+                raise ValueError("DOL 公告时间与所选日程不符，未填入其他期次的数据")
+            return response, raw_id
         if row.source in europe.SCHEDULES:
             return await self.european_result(row)
         date = row.scheduled_at.astimezone(EASTERN).date()
@@ -275,7 +310,13 @@ class CalendarWorker:
                 page = await self.result_page(row)
                 if page is not None:
                     response, raw_id = page
-                    if row.source in europe.SCHEDULES:
+                    if row.source == "eia":
+                        announcement = eia.report(response.text, self.eia_schedule_body)
+                        period, stage, values = announcement.period, announcement.stage, announcement.values
+                    elif row.source == "dol":
+                        announcement = (await asyncio.to_thread(dol.report, await asyncio.to_thread(pdf_text, response.content)))[row.family]
+                        period, stage, values = announcement.period, announcement.stage, announcement.values
+                    elif row.source in europe.SCHEDULES:
                         announcement = europe.parse_announcement(row.family, response.text)
                         period, stage, values = announcement.period, announcement.stage, announcement.values
                     elif row.family == "umich":