feat(playback): pre-transcode HEVC videos in background; veryfast preset

Long videos blocked /playback for the entire encode duration. The fix
is to populate the cache before the user clicks, not when they click.

Changes:
- Extract ffprobe + ffmpeg helpers to services/video.py so the request
  handler and the background task share one sync implementation. The
  endpoint wraps calls in asyncio.to_thread; celery just calls them.
- New tasks/video.py with pretranscode_video. Idempotent: skips when
  the cache is already current and skips passthrough-safe sources
  (h264 in mp4/m4v/webm). 30-min task time limit so the long-tail
  files (3GP archive, multi-minute 1080p clips) still complete.
- scan_folder now dispatches pretranscode_video alongside
  generate_thumbnails / extract_metadata for any new video row.
- POST /library/maintenance/backfill-video-cache enqueues every
  active video so the existing library catches up.
- libx264 preset bumped from fast to veryfast. ~2x throughput on this
  CPU-only box, output a few % larger but well within disk budget.
- /playback simplifies to: cache check, passthrough if h264 in
  web-safe container, else sync transcode (still there as fallback
  for races against the queued task).

Once the backfill task drains, /playback should be near-instant for
every video. Any video added afterwards is pre-transcoded at scan
time, so the user keeps that property going forward.

backend/app/tasks/video.py

@@ -0,0 +1,62 @@
+"""Background pre-transcode for video photos so /playback is a cache
+hit on first user click. Dispatched from scan_folder when a new video
+row is created, and from the backfill admin endpoint for the existing
+library."""
+from __future__ import annotations
+
+import logging
+import os
+
+from celery import shared_task
+
+from app.services.video import (
+    PLAYBACK_OK_EXTS,
+    PLAYBACK_OK_VCODECS,
+    cache_path_for,
+    ffprobe_video_codec,
+    transcode_to_h264_mp4,
+)
+
+
+logger = logging.getLogger(__name__)
+
+
+@shared_task(
+    name='pretranscode_video',
+    # Override the 10-minute global task_time_limit. Long videos (30+ min
+    # raw clips, the occasional .3gp from the 2016 archive) can legitimately
+    # take half an hour to transcode on this CPU-only box.
+    time_limit=1800,
+    soft_time_limit=1740,
+)
+def pretranscode_video(photo_id: str, src_path: str):
+    """Idempotent: skip if cache exists and is newer than source, skip
+    if source is already a passthrough-safe codec/container.
+
+    Reports the outcome as a status string so the admin backfill can
+    summarise. The /playback endpoint also falls back to a sync
+    transcode if the cache miss races a queued task."""
+    if not os.path.exists(src_path):
+        logger.debug("pretranscode skip — source missing: %s", src_path)
+        return {'status': 'missing', 'photo_id': photo_id}
+
+    cache_path = cache_path_for(photo_id)
+    if cache_path.exists():
+        try:
+            if os.path.getmtime(src_path) <= os.path.getmtime(cache_path):
+                return {'status': 'cached', 'photo_id': photo_id}
+            cache_path.unlink()
+        except OSError:
+            pass
+
+    ext = os.path.splitext(src_path)[1].lower()
+    if ext in PLAYBACK_OK_EXTS:
+        codec = ffprobe_video_codec(src_path)
+        if codec in PLAYBACK_OK_VCODECS:
+            return {'status': 'passthrough', 'photo_id': photo_id, 'codec': codec}
+
+    ok = transcode_to_h264_mp4(src_path, str(cache_path))
+    return {
+        'status': 'transcoded' if ok else 'failed',
+        'photo_id': photo_id,
+    }