feat: runtime feature flags, upload/download, RAW decoding

Adds Redis-backed feature flags for vision stages with admin UI toggles
and manual backfill trigger, photo upload and download routers with
frontend upload modal, and rawpy-based RAW decoding with JPEG fallback
for misnamed DNGs. Fixes pgvector serialization, is_trashed filter, and
naive-datetime bind in incremental duplicate regrouping; bumps Celery
time limits on regroup tasks beyond the 5-minute default.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

backend/app/services/duplicates.py

@@ -221,6 +221,13 @@ async def incremental_regroup(
             from datetime import timedelta
             since = datetime.now(timezone.utc) - timedelta(hours=1)
 
+        # Photo.added_at is stored as TIMESTAMP WITHOUT TIME ZONE, so
+        # asyncpg rejects aware datetimes with "can't subtract offset-naive
+        # and offset-aware". Normalise: if `since` has a tzinfo, convert
+        # it to UTC and drop the tzinfo so the bind parameter is naive.
+        if since.tzinfo is not None:
+            since = since.astimezone(timezone.utc).replace(tzinfo=None)
+
         # Get newly added photos (the "new" set).
         new_rows = (
             await session.execute(
@@ -349,6 +356,14 @@ async def _clip_neighbor_scan(
     for photo_id, vector in target_embeddings:
         # pgvector cosine distance: <=> operator
         # Find top 20 nearest neighbors within threshold.
+        # Serialize the vector as "[a,b,c,...]" — pgvector's text
+        # format uses commas; numpy's default str() joins with spaces
+        # which Postgres rejects with "invalid input syntax for vector".
+        if hasattr(vector, 'tolist'):
+            vec_seq = vector.tolist()
+        else:
+            vec_seq = list(vector)
+        vec_text = '[' + ','.join(f'{float(x):.8f}' for x in vec_seq) + ']'
         result = await session.execute(
             text("""
                 SELECT e.photo_id, (e.vector <=> :vec) AS distance
@@ -356,14 +371,14 @@ async def _clip_neighbor_scan(
                 JOIN photos p ON p.id = e.photo_id
                 WHERE e.model = :model
                   AND e.photo_id != :pid
-                  AND p.is_discarded = false
+                  AND p.is_trashed = false
                   AND p.is_hidden = false
                   AND (e.vector <=> :vec) < :threshold
                 ORDER BY e.vector <=> :vec
                 LIMIT 20
             """),
             {
-                'vec': str(vector),
+                'vec': vec_text,
                 'pid': photo_id,
                 'model': embedder_model,
                 'threshold': threshold,