Follow-up to slice 3: a read-only artist preview page keyed by MBID, reachable from
suggested artists, suggested albums, and Find-similar results. Shows discography,
lazy per-album tracklists, own-state badges; Follow (reuse /api/artists) + Want
(new /api/discover/want upsert). Web-side MB reads only, no worker changes.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
README: discovery status → done (ListenBrainz-native, no Spotify); off-by-default
note for the discovery sweep; architecture mentions MusicBrainz+ListenBrainz.
Plan: check off final verification (suites green, whole-branch review passed,
live-API algorithm fix), append the review's deferred items to Notes.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The placeholder algorithm string (contribution_5_threshold_15_limit_50) is not a
permitted enum member of the LB Labs endpoint — every lookup 400'd, making discovery
inert. Verified against the live API 2026-07-12: the days_7500/contribution_3/
threshold_10/limit_100 variant returns rows shaped {artist_mbid,name,comment,score},
which the existing defensive parser already handles. Fixed in both the worker adapter
and the web seed-search client (kept in sync).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Adds registry.build_similarity_sources() (always constructs
ListenBrainzSource; reachability is a per-run health() concern, not a
startup gate) and wires discovery into the worker's main loop: a
scheduled sweep gated on discover.enabled + DISCOVER_TICK_SECONDS, and
a one-shot discover.requested trigger mirroring the existing
scan.requested pattern.
Also hardens run_discovery's health-check pass: a source whose
health() raises is now logged and skipped instead of aborting the
whole sweep (accepted finding from the Task-3 review).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Spotify similarity/new-release APIs are dead for new apps (Nov 2024 + Feb 2026),
so discovery is built on ListenBrainz similar-artists (MBID-native) behind a
pluggable SimilaritySource interface. Manual review feed of similar artists +
albums; Follow/Want reuse slice-2 operations, Dismiss is permanent. Off by default.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
MonitoredRelease rows created before an artist's discography was
populated (e.g. an owned album recorded before _populate_discography
ran) never got primaryType/secondaryTypes backfilled, wrongly
excluding them from the studio-only core view. _populate_discography
now DO UPDATEs the MB type metadata on conflict instead of DO NOTHING,
leaving monitored/state/currentQualityClass/firstGrabbedAt untouched.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Adds the per-artist release-type filter toggle: off (default) restricts
a discography view to core studio releases, on shows everything
(live, compilations, etc.).
A followed artist previously showed only the one owned album ("of 1")
because scan_library recorded just the resolved release. Now scan
browses each artist's full MusicBrainz release-group list once per
run and stores it as unmonitored MonitoredRelease rows, matching
web-follow behavior. The owned album keeps its monitored/fulfilled
state via ON CONFLICT DO NOTHING.
Type the POST request param (_request: Request) to match the test's
call signature and keep tsc --noEmit clean, per Next.js route handler
convention already used elsewhere in the codebase.
Probe each album exactly once and thread the ProbeResult through for
tags, quality, and format instead of hardcoding LibraryItem.format to
"FLAC" and re-probing on the tag-fallback path. Also skip album-level
directories starting with "." (previously only artist-level dirs were
skipped).
Adds the AudioProbe seam (probe.py, _probe.py MutagenProbe with a lazy
mutagen import, FakeAudioProbe) and scan_library, which walks /music,
resolves each album against MusicBrainz (falling back to embedded tags
when the folder name doesn't resolve), and idempotently records matched
albums as have (LibraryItem), monitored (MonitoredRelease, fulfilled),
and followed (WatchedArtist, autoMonitorFuture=false).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Test suites were running DELETE FROM.../deleteMany() against whatever
DATABASE_URL pointed to, which was the live app database, wiping user
credentials and library. Add a guard in both the worker conftest and
web test setup that raises before any cleanup unless the target DB
name ends in '_test'. Default DATABASE_URL to lyra_test when unset in
the worker, and make `npm test` target lyra_test explicitly.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Rewrite run_pipeline's download->import section to download into a per-job
staging dir (cleared before each candidate), verify completeness there,
promote only a complete album into the library via import_album, tag the
final dir, then DB-import. The staging dir is always removed in a
try/finally, so a partial/failed download never touches or corrupts the
library, and a retry can no longer leave orphan files behind.
Also point test_upgrade_e2e.py's run_pipeline call at dest_root="/tmp/lib"
(matching every other pipeline test) since import_album now unconditionally
creates the final dir, which requires a writable dest_root.
The worker's enqueue_due refuses to enqueue when a non-terminal job
already exists for a release. The search-now route enqueued
unconditionally, so two clicks (or a click racing the worker sweep)
could create two concurrent pipelines for the same album, both
importing and colliding on LibraryItem's @@unique([artist,album]).
Mirror the worker's guard: no-op if a job in a non-terminal state
(anything other than imported/needs_attention) already exists for the
release.
Jonathan