fix(workers): worker-level recovery — autoretry on transient errors + tightened sweep (5min) + import_media_file body-wrap so no path leaves rows stuck in 'processing'

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

backend/app/tasks/ml.py

@@ -9,6 +9,7 @@ apply_allowlist_tags sweeps which are 'maintenance' lane. Sync sessions
 from pathlib import Path
 
 from sqlalchemy import select
+from sqlalchemy.exc import DBAPIError, OperationalError
 
 from ..celery_app import celery
 from ..models import ImageRecord, MLSettings
@@ -22,7 +23,17 @@ def _is_video(path: Path) -> bool:
     return path.suffix.lower() in VIDEO_EXTS
 
 
-@celery.task(name="backend.app.tasks.ml.tag_and_embed", bind=True)
+@celery.task(
+    name="backend.app.tasks.ml.tag_and_embed",
+    bind=True,
+    autoretry_for=(OperationalError, DBAPIError, OSError),
+    retry_backoff=5,
+    retry_backoff_max=60,
+    retry_jitter=True,
+    max_retries=3,
+    soft_time_limit=300,
+    time_limit=420,
+)
 def tag_and_embed(self, image_id: int) -> dict:
     """Run Camie + SigLIP on one image; store predictions + embedding;
     then enqueue per-image allowlist application.