feat: M2 (part 2) — RSS poller, backfill, enrichment, scheduler

- Free per-channel RSS reader for quota-less fresh-video detection
- Recent-first backfill (configurable: 100 videos / 1 year) plus resumable deep
  backfill from the uploads playlist
- Enrichment via videos.list: duration, view/like counts, category, topics,
  language, Shorts heuristic and livestream/premiere classification
- Reusable sync runners + APScheduler jobs (rss / enrich / backfill), all
  quota-aware with a reserve so backfill never starves fresh enrichment
- Manual triggers: POST /api/sync/{rss,backfill,enrich}
- Exact insert counting via RETURNING with in-batch de-duplication

backend/app/scheduler.py

@@ -0,0 +1,77 @@
+"""Background scheduler: free RSS detection, enrichment of new videos, and quota-aware
+recent-first / deep backfill. Runs inside the API process (single worker)."""
+import logging
+
+from apscheduler.schedulers.background import BackgroundScheduler
+
+from app.config import settings
+from app.db import SessionLocal
+from app.sync.runner import (
+    run_deep_backfill,
+    run_enrich,
+    run_recent_backfill,
+    run_rss_poll,
+)
+
+logger = logging.getLogger("subfeed.scheduler")
+
+_scheduler: BackgroundScheduler | None = None
+
+
+def _job(name: str, fn) -> None:
+    db = SessionLocal()
+    try:
+        result = fn(db)
+        logger.info("job %s -> %s", name, result)
+    except Exception:
+        db.rollback()
+        logger.exception("job %s failed", name)
+    finally:
+        db.close()
+
+
+def _rss_job() -> None:
+    _job("rss_poll", lambda db: run_rss_poll(db))
+
+
+def _enrich_job() -> None:
+    _job("enrich", lambda db: run_enrich(db))
+
+
+def _backfill_job() -> None:
+    # Recent-first for not-yet-synced channels, then deep backfill for the rest.
+    def work(db):
+        recent = run_recent_backfill(db, max_channels=25)
+        deep = run_deep_backfill(db, max_channels=10)
+        return {"recent": recent, "deep": deep}
+
+    _job("backfill", work)
+
+
+def start_scheduler() -> None:
+    global _scheduler
+    if not settings.scheduler_enabled or _scheduler is not None:
+        return
+    scheduler = BackgroundScheduler(timezone="UTC")
+    scheduler.add_job(
+        _rss_job, "interval", minutes=settings.rss_poll_minutes, id="rss_poll"
+    )
+    scheduler.add_job(
+        _enrich_job, "interval", minutes=settings.enrich_interval_minutes, id="enrich"
+    )
+    scheduler.add_job(
+        _backfill_job,
+        "interval",
+        minutes=settings.backfill_interval_minutes,
+        id="backfill",
+    )
+    scheduler.start()
+    _scheduler = scheduler
+    logger.info("scheduler started")
+
+
+def shutdown_scheduler() -> None:
+    global _scheduler
+    if _scheduler is not None:
+        _scheduler.shutdown(wait=False)
+        _scheduler = None