Fix stuck dedup candidate, enrich timeout/starvation, and qobuz status error

1. enrich-buy-url.py re-queried every no-match track on every daily run (external,
   rate-limited lookups over the whole backlog), which blew past its 30-min
   timeout. Cache "tried, no match" with a BUY_URL_TRIED marker tag and skip
   those for 30 days (--force still re-checks). This is what made "Look up buy
   links" fail with TIMEOUT after 1800s.

2. Move maintenance:enrich_buy_url to run last (09:45) in the 9am block, after
   the fingerprint index (09:25) and status report (09:30). enrich starting at
   09:10 and holding the shared lock is what left "Rebuild fingerprint index"
   skipped_lock.

3. pipeline-status.sh: guard the qobuz/app_id read with a file-exists check.
   `< missing 2>/dev/null` still leaks the shell's redirection error (opened
   before 2>/dev/null applies), so the daily report logged
   "qobuz/app_id: No such file or directory" every run.

4. dedup confirm_and_apply: a candidate whose delete target no longer exists
   (already removed by an earlier dedup/upgrade/hand) was filtered out and the
   apply silently no-op'd, leaving it stuck in the queue with no way to delete
   or clear it (the 100 gecs case). Now mark such candidates applied so a Delete
   click clears them. Covered by tests/test_dedup_review.py.

Tests: 46 green.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

app/services/dedup_review_service.py

@@ -208,15 +208,33 @@ async def confirm_and_apply(candidate_ids: list[int], confirmed_by: str) -> Dedu
     try:
         candidates = [db.get(DedupCandidate, cid) for cid in candidate_ids]
         candidates = [c for c in candidates if c is not None and not c.applied]
-        candidates = [c for c in candidates if Path(c.delete_path).exists() and Path(c.keep_path).exists()]
         if not candidates:
             return None
 
+        now = time.time()
+
+        # A candidate whose delete target no longer exists (already removed by an
+        # earlier dedup, the mp3->flac upgrade, or by hand) is effectively already
+        # resolved. Mark it applied so it clears from the pending queue, instead
+        # of the old behavior of filtering it out and silently no-oping -- which
+        # left such a stale candidate stuck in the list forever with no way to
+        # delete or clear it.
+        already_gone = [c for c in candidates if not Path(c.delete_path).exists()]
+        for c in already_gone:
+            c.confirmed = True
+            c.confirmed_by = confirmed_by
+            c.confirmed_at = now
+            c.applied = True
+        if already_gone:
+            db.commit()
+
+        # The rest need both files present for a safe delete.
+        actionable = [c for c in candidates if Path(c.delete_path).exists() and Path(c.keep_path).exists()]
+
         by_script: dict[str, list[DedupCandidate]] = {}
-        for c in candidates:
+        for c in actionable:
             by_script.setdefault(_script_for_pass(c.pass_name), []).append(c)
 
-        now = time.time()
         finished_at = now
         log_paths = []
         ran_candidates: list[DedupCandidate] = []
@@ -246,7 +264,8 @@ async def confirm_and_apply(candidate_ids: list[int], confirmed_by: str) -> Dedu
             ran_candidates.extend(group)
         db.commit()
 
-        if not ran_candidates:
+        # Nothing to report: no stale ones cleared and nothing ran.
+        if not already_gone and not ran_candidates:
             return None
 
         still_there = {c.delete_path for c in ran_candidates if Path(c.delete_path).exists()}
@@ -261,7 +280,7 @@ async def confirm_and_apply(candidate_ids: list[int], confirmed_by: str) -> Dedu
             started_at=now,
             finished_at=finished_at,
             mode="apply",
-            deleted=actually_deleted,
+            deleted=actually_deleted + len(already_gone),
             kept=len(ran_candidates) - actually_deleted,
             log_path=";".join(log_paths) or None,
         )

app/services/scheduler_service.py

@@ -130,17 +130,19 @@ MAINTENANCE_JOBS: dict[str, tuple[dict, callable]] = {
         dict(minute=5, hour=9),
         _lib("maintenance:export_laptop_playlists", "export-laptop-playlists.py"),
     ),
-    "maintenance:enrich_buy_url": (
-        dict(minute=10, hour=9),
-        # 30-min cap: this hits external buy-link APIs per track and has hung
-        # holding the pipeline lock (2026-07-09/10), starving every job after
-        # it. A timeout releases the lock so the rest of the 9am chain runs.
-        _lib("maintenance:enrich_buy_url", "enrich-buy-url.py", ["--apply"], timeout=1800),
-    ),
     "maintenance:build_fingerprint_index": (
         dict(minute=25, hour=9),
         _lib("maintenance:build_fingerprint_index", "build-fingerprint-index.py", ["--workers", "8"], timeout=3600),
     ),
+    "maintenance:enrich_buy_url": (
+        # Runs LAST in the 9am block: it hits external buy-link APIs per track
+        # and can run long, so scheduling it after the fingerprint index (09:25)
+        # and the status report (09:30) keeps it from starving them via the
+        # shared lock. The 30-min cap is a backstop; enrich-buy-url.py also now
+        # caches "tried, no match" so it stops re-querying the whole backlog.
+        dict(minute=45, hour=9),
+        _lib("maintenance:enrich_buy_url", "enrich-buy-url.py", ["--apply"], timeout=1800),
+    ),
     "maintenance:pipeline_status_report": (
         dict(minute=30, hour=9),
         # Read-only (reads logs, pings Navidrome, sends Telegram). Runs WITHOUT