Commit Graph
5 Commits
Author SHA1 Message Date
mrleo1nid 02307802bc style(recommender): format demographic integration test
ci / recommender / recommender (push) Successful in 37s
release / release (версия и заметки) (push) Successful in 5s
release / prune (старые файлы и прогоны) (push) Successful in 5s
ci / db / db (push) Successful in 54s
release / android (APK в релиз) (push) Successful in 3m54s
release / desktop (Linux и Windows в релиз) (push) Successful in 12m38s
2026-10-22 14:25:11 +03:00
mrleo1nid fff91c446b feat(recommendations): collect demographics and advance age cohorts
ci / ts / ts (packages + web) (push) Successful in 3m19s
ci / db / db (push) Successful in 53s
ci / module boundaries / boundaries (push) Successful in 7s
ci / api / api (push) Successful in 51s
ci / contract / contract (сгенерированный код совпадает с openapi.yaml) (push) Successful in 1m3s
ci / recommender / recommender (push) Failing after 16s
ci / mobile / mobile (Metro-бандл Android) (push) Successful in 1m15s
2026-10-22 14:19:52 +03:00
mrleo1nidandClaude Opus 5.5 58bd14ef38 feat(recommender): LightGBM lambdarank ranker with impression features (M6)
ci / db / db (push) Failing after 1m15s
ci / api / api (push) Successful in 17s
ci / recommender / recommender (push) Successful in 32s
ci / docker / docker (образы собираются — их сервер собирает сам при обновлении) (push) Successful in 1m41s
The wave logs each served candidate's feature vector to rec_impression_features.
Nightly training joins them with impression outcomes, validates on the last week
and activates the model only when it beats the v1 order and the active model.
The service loads the active ranker and uses its order in place of the v1 score.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-20 22:41:26 +03:00
mrleo1nidandClaude Opus 5.5 fb07f962eb feat: taste onboarding for new listeners (M6)
ci / contract / contract (сгенерированный код совпадает с openapi.yaml) (push) Successful in 1m2s
ci / db / db (push) Failing after 1m29s
ci / ts / ts (packages + web) (push) Successful in 49s
ci / api / api (push) Successful in 1m25s
ci / mobile / mobile (Metro-бандл Android) (push) Successful in 48s
ci / recommender / recommender (push) Successful in 21s
- GET /onboarding offers the library's top artists (with album covers)
  and genres and says whether to show it (not done and < 10 plays);
  PUT /me/onboarding stores the choice or a skip (migration 00009).
- The recommender seeds the taste from it while the user has fewer than
  30 scored tracks: chosen artists' tracks +1, chosen genres +0.5.
- Web: onboarding replaces the wave on the home page until done or
  skipped; artist grid with covers, genre chips, at least 3 artists.
- e2e journey and the recommender DB test cover it.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-20 22:35:43 +03:00
mrleo1nidandClaude Opus 5.5 f5c276092f ci: integration tests on PostgreSQL + pgvector for migrations and store SQL
ci / db / db (push) Failing after 53s
ci / recommender / recommender (push) Successful in 36s
ci / api / api (push) Successful in 34s
ci / docker / docker (образы собираются — их сервер собирает сам при обновлении) (push) Successful in 1m9s
New `ci / db` job with a pgvector service: Go applies all migrations,
rolls them back to zero and applies them again, then runs the scanner,
fingerprint, write-template, top-tracks and upload-queue store queries;
Python runs the analysis worker, catalog / profile loading and a full
ALS training run against the same schema. Until now this SQL ran
nowhere before reaching a real server.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-20 18:18:44 +03:00