feat(worker): share the MusicBrainz cache from scans/discovery

The worker's MB browser (discography + artist search, used by monitor sweeps,
library scans, and Last.fm discovery) re-fetched every time. Route it through
the same ApiCache table + logical keys the web app uses, so a discography one
process fetched warms the other, and repeat lookups stop hitting MB.

- apicache.py cached_api(): read-through, serve-stale-on-outage, never caches
  null, best-effort (a cache-DB error degrades to a live fetch, never breaks
  scan/sweep/discovery). Its own DEDICATED autocommit connection so it can't
  commit the worker's in-flight transaction, and autocommit avoids freezing
  Postgres now() (which would break TTL math).
- CachingMbBrowser decorates MusicBrainzBrowser, serializing to the SAME
  camelCase JSON shape the web writes (round-trip tested). registry.build_browser
  wraps it when a DSN is present; the Last.fm source's browser rides it too.
- Hourly prune drops ApiCache rows > 90d so the table stays bounded.
- worker 213 tests / 7-skip.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

worker/lyra_worker/main.py

@@ -23,6 +23,7 @@ IDLE_SLEEP = 2.0
 HEARTBEAT_SECONDS = 15.0  # how often to stamp worker.heartbeat (its updatedAt = liveness)
 MONITOR_TICK_SECONDS = 60.0
 DISCOVER_TICK_SECONDS = 300.0  # check the discovery interval every 5 min (interval itself is hours)
+CACHE_PRUNE_SECONDS = 3600.0  # prune the shared ApiCache at most hourly (drops rows > 90d old)
 DEST_ROOT = "/music"  # must match run_pipeline's default library root
 # Per-job download staging. Defaults inside the library (/music/.staging) to preserve prior
 # behavior; set STAGING_ROOT (mounted as a separate volume) to keep temp download I/O off a
@@ -167,6 +168,23 @@ def maybe_run_discovery(conn, sources, browser, config, dcfg, now, last_discover
     return now if due else last_discover_tick
 
 
+def prune_api_cache(conn) -> None:
+    """Drop ApiCache rows older than 90 days so the shared cache stays bounded. Best-effort:
+    a failure here (e.g. a transient error) must never take down the worker loop."""
+    try:
+        with conn.cursor() as cur:
+            cur.execute('DELETE FROM "ApiCache" WHERE "fetchedAt" < now() - interval \'90 days\'')
+        conn.commit()
+    except psycopg.OperationalError:
+        raise  # a real disconnect — let the loop's reconnect handler deal with it
+    except Exception as e:
+        print(f"worker: ApiCache prune failed: {e}", flush=True)
+        try:
+            conn.rollback()
+        except Exception:
+            pass
+
+
 def _require_secret_key() -> None:
     """Fail fast (or warn loudly) on a missing/placeholder/public LYRA_SECRET_KEY before
     the worker touches any encrypted credential."""
@@ -192,9 +210,10 @@ def run_forever() -> None:
     adapters = build_adapters(get_config(conn))
     resolver = build_resolver()
     tagger = build_tagger()
-    browser = build_browser()
+    _dsn = os.environ.get("DATABASE_URL")
+    browser = build_browser(_dsn)  # shared ApiCache read-through when a DSN is present
     probe = build_probe()
-    similarity_sources = build_similarity_sources(get_config(conn))
+    similarity_sources = build_similarity_sources(get_config(conn), _dsn)
     # Backfill: releases pressed while the monitor was off never got reconciled, so they linger
     # on the wanted list. Fix any already-owned ones once at startup.
     fixed = fulfill_owned_releases(conn, MonitorConfig.from_config(get_config(conn)))
@@ -205,6 +224,7 @@ def run_forever() -> None:
     last_tick = 0.0
     last_discover_tick = 0.0
     last_heartbeat = 0.0
+    last_prune = 0.0
     try:
         while True:
             # A transient DB connection loss (e.g. AdminShutdown when Postgres restarts) raises
@@ -218,6 +238,9 @@ def run_forever() -> None:
                 if now - last_heartbeat >= HEARTBEAT_SECONDS:
                     _set_config(conn, "worker.heartbeat", "1")  # value irrelevant; updatedAt is the signal
                     last_heartbeat = now
+                if now - last_prune >= CACHE_PRUNE_SECONDS:
+                    prune_api_cache(conn)  # keep the shared ApiCache bounded
+                    last_prune = now
                 if mcfg.enabled and now - last_tick >= MONITOR_TICK_SECONDS:
                     try:
                         sweep(conn, browser, mcfg)