- Deezer: /artist/{id}/radio for collection artists, affinity 2.0
- YouTube Music: RDAMVM watch playlist via yt-dlp, affinity 2.0
- seed radios from the 2 newest songs of the mood's own catalog
- dedupe candidates by recording, not only by external id
- clean raw YouTube titles and Topic suffixes when indexing
- tests, smoke marker for the RDAMVM path, docs
- seed from every artist (weighted by song count) and every album, not one track
- walk Deezer's graph: artist top tracks, related artists, album deep cuts
- score by artist frequency, collection album, catalog affinity and hint text
- penalise repeated artists so a 30-slot mood spans artists, not one
- refresh the pool when a song is added instead of waiting for refreshHours
- skip discovery for a mood with no hint, genres or songs
- extend tests and the smoke test with the artist graph
- V9 adds source/external_id to songs, moods, recommendations and jobs
- Deezer public catalog for search, related album covers and metadata
- streamrip downloader configured with DEEZER_ARL, 128/320 MP3
- per-mood source selector; switching clears and refills recommendations
- keep YouTube Music as a full second catalog and downloader
- media layout <source>-<externalId>, legacy YouTube folders still read
- Docker/streamrip install, chart ARL secret, docs and dual-source smoke test
- V8 wipe migration and the shared song store keyed by YouTube video id
- media store, indexer, downloader, recommendation top-up and refcounted GC
- mood songs plus the recommendation lifecycle and mixed radio pools
- new search, song and recommendation API; remove library/longform/seed routes
- chart PVC/media env, docs, Nix, README and an end-to-end smoke test
Scanner and library:
- Keep the last library when a configured folder cannot be walked; the
previous code deleted every song, cascading away mood seeds and progress.
- Re-adding a file now clears its tombstone, and a re-deletion refreshes the
timestamp instead of keeping the old one.
- Tombstone timestamps are unique per row, and delta pages continue from the
older of the song/tombstone page ends.
- Snapshot pagination accepts a stable revision for later pages so songs
inserted mid-sync are not skipped forever.
- Reject "%" and "_" searches that normalize to an empty LIKE needle.
Downloads and discovery:
- Download promotion uses the deterministic path id, one folder per track so
covers do not overwrite each other, removes failed promotions, and restarts
the scan after a running one finishes.
- Stale downloading jobs are recovered; reject cancels queued/in-flight jobs;
markDone only owns downloading rows; rejecting no longer seeds the track.
- Reuse an already completed download of the same video for another mood.
- yt-dlp stdout and stderr are separated, empty discovery results no longer
prune undecided recommendations, and filenames are capped by UTF-8 bytes.
API and data:
- Malformed or missing JSON returns 400 instead of 500; optional session
bodies use receiveNullable.
- Playback events must belong to the session, known song and open session.
- Radio refills the candidate pool after recent-song pruning.
- Seed percent is required in MoodRepository.update.
- Normalize legacy V4 liked/downloaded recommendation states before V5, with
an upgrade regression test.
CI: check out without persisting credentials and require tests before chart
publishing.
yt-dlp now feeds mood discovery. Queries come from the free-text hint, seed
artists and genres; they run in parallel, are filtered to music-length videos,
matched against library songs by normalized title and artist, scored, and
stored as review candidates. Refresh, accept and reject endpoints are added:
accepting a library match adds a seed immediately, while accepted tracks that
are not owned wait for the future downloader. Decisions survive refreshes and
stale undecided candidates are pruned.
A background pass refreshes stale moods every hour, and the image ships a
pinned yt-dlp binary (YTDLP_PATH, also added to the Nix shell and chart).
Radio now avoids tracks played in the last 7 days when the pool is large
enough, spaces artists and albums across a batch, and weighs skips three times
as heavily as completions. The chart moves to 0.2.2.