danvics/soulsync
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 1
soulsync/tests
History
Broque Thomas 402d851cac Deezer search: drop advanced-syntax at endpoint, free-text + rerank wins 
Live-API verification revealed advanced-syntax queries hurt more
than they help on this endpoint. Switching the import-modal Deezer
search back to free-text + local rerank.

# What live testing showed

Hit Deezer's public API with both query forms for the issue #534
case (`Dirty White Boy` + `Foreigner`):

**Free-text (`q=Dirty White Boy Foreigner`):**
- Returns 21 results
- Real Foreigner Head Games studio cut at #1
- Live versions at #2-10
- Karaoke / cover variants at #11-15

**Advanced (`q=track:"Dirty White Boy" artist:"Foreigner"`):**
- Returns 12 results
- "(2008 Remaster)" at #1 — canonical Head Games cut MISSING from
  top 8 entirely
- Live + alt-album versions follow

Advanced syntax DOES filter karaoke at the API level (none in the
12-result set vs. 5 at positions 11-15 in free-text), but it has
its own ranking bias that surfaces remasters / "Best Of" cuts
ahead of the canonical recording. Net regression for the user-
facing goal.

# Fix

1. Endpoint reverts to free-text query with local rerank applied.
2. Local rerank gains "remaster" / "remastered" / "reissue"
   patterns under VARIANT_TAG_PATTERNS (soft 0.4× penalty — user
   may want them but they shouldn't outrank the original).
3. Client kwarg support (`track=` / `artist=` / `album=`) preserved
   for future opt-in callers (e.g. exact-match flows where API-
   level filtering matters more than ranking).

# Verified end-to-end against live Deezer API

Re-ran the exact #534 case through the live API + new rerank.
Top 15 results post-rerank:

1. Dirty White Boy — Foreigner — Head Games  ← REAL CUT AT TOP
2-10. Various Live versions
11-15. Karaoke / cover / tribute variants  ← BURIED

Real Foreigner Head Games studio cut at #1, exactly the user's
ask.

# Tests

- `test_relevance.py` — variant tag patterns extended; existing
  tests still pass (50 tests).
- `test_search_match_endpoints.py::test_joins_track_and_artist_into_free_text_query`
  — replaces `test_passes_track_and_artist_as_kwargs`; verifies
  endpoint sends free-text join, NOT field-scoped kwargs (the
  prior test asserted the wrong direction now).
- Karaoke-burying assertion at the endpoint still pins the
  user-visible behaviour.
- Client kwarg path tests untouched (still pin advanced-syntax
  construction for future opt-in callers).

# Verification

- 75 relevance + endpoint + query tests pass
- 2445 full suite passes
- Ruff clean
- Live Deezer API shows real cut at #1 post-rerank
2026-05-10 09:36:48 -07:00
..
artists Direct ID lookup in Enhance Quality, like Download Discography 2026-05-06 12:05:41 -07:00
automation Surface handler-returned errors in automation last_error 2026-04-30 15:45:28 -07:00
discovery MS Cin-5: Drop per-server globals — engine owns the clients 2026-05-05 18:05:45 -07:00
downloads Bound slskd HTTP timeout — fixes worker thread deadlock 2026-05-06 22:02:25 -07:00
enrichment Honor manually-matched source IDs in per-source enrichment workers 2026-05-06 19:00:53 -07:00
imports Deezer search: drop advanced-syntax at endpoint, free-text + rerank wins 2026-05-10 09:36:48 -07:00
library Fix Album Completeness Auto-Fill on Docker / shared-library setups (#476) 2026-05-03 10:11:06 -07:00
media_server Add cross-section dedup to all-libraries listing layer 2026-05-06 16:01:37 -07:00
metadata Deezer search: free-text fallback when advanced query returns 0 2026-05-10 09:16:13 -07:00
playlists Lift _prepare_stream_task + playlist_explorer_build_tree to core/ 2026-04-29 11:26:07 -07:00
search Cin-6: Rename soulseek_client global → download_orchestrator 2026-05-04 23:23:32 -07:00
static Discover controller — Cin pre-review polish 2026-05-07 20:35:10 -07:00
stats Lift /api/stats/* and /api/listening-stats/* into core/stats/ 2026-04-27 14:27:03 -07:00
streaming Cin-6: Rename soulseek_client global → download_orchestrator 2026-05-04 23:23:32 -07:00
watchlist Lift _process_watchlist_scan_automatically to core/watchlist/auto_scan.py 2026-04-29 08:17:41 -07:00
wishlist Coerce wishlist track counts before category checks 2026-04-30 21:42:43 +03:00
conftest.py Split metadata source and Spotify status 2026-05-02 22:02:00 +03:00
test_acoustid_scanner.py Fix null-pointer error in acoustid_scanner 2026-04-17 19:17:24 +03:00
test_acoustid_skip_logic.py Fix: tasks showed Completed when file was quarantined 2026-05-03 18:28:32 -07:00
test_acoustid_version_mismatch.py Reject AcoustID matches whose version disagrees with the expected track 2026-05-07 13:25:30 -07:00
test_album_completeness_job.py Fix: maintenance findings badge inflated when scan dedup-skipped 2026-05-04 08:55:13 -07:00
test_album_mbid_consistency.py Update mbid consistency test mock to match new create_finding bool contract 2026-05-04 09:11:01 -07:00
test_artist_top_tracks_clients.py Add download buttons + bulk action to artist top-tracks sidebar 2026-05-07 15:44:47 -07:00
test_auth_usage_throttle.py test: auth last_used_at write throttle 2026-04-19 15:22:24 -07:00
test_config_save_retry.py Stop config retry tests from writing to the real DB 2026-04-30 14:07:00 -07:00
test_content_filter_regex.py Fix watchlist content filters: live false positives + auto-scan bypass 2026-04-21 19:16:25 -07:00
test_create_finding_dedup_counter.py Repair job card badge — show pending count, not last-scan count 2026-05-07 08:28:17 -07:00
test_dashboard_events.py Realign test file names 2026-04-17 21:25:01 +03:00
test_discogs_collection_source.py Foundation: typed Album/Track/Artist + per-provider converters 2026-05-03 22:21:32 -07:00
test_discogs_track_count.py Discogs: count rows with empty type_ as real tracks too 2026-04-25 08:27:44 -07:00
test_discover_section_controller_js.py Discover controller — Cin pre-review polish 2026-05-07 20:35:10 -07:00
test_download_orchestrator_soundcloud.py Cin-5: Drop per-source attrs from orchestrator 2026-05-04 23:14:05 -07:00
test_download_plugin_conformance.py Cin-5: Drop per-source attrs from orchestrator 2026-05-04 23:14:05 -07:00
test_downloads_pagination.py Lift shared runtime state into core 2026-04-27 19:54:44 +03:00
test_duplicate_detector_mount_paths.py Filter same-physical-file duplicates from duplicate detector 2026-04-30 14:06:10 -07:00
test_duplicate_detector_slskd_dedup.py Fix: maintenance findings badge inflated when scan dedup-skipped 2026-05-04 08:55:13 -07:00
test_enrichment_services.py Consolidate enrichment bubble routes behind a service registry 2026-05-01 20:14:30 -07:00
test_featured_artist_completion.py Match featured-artist tracks across discography completion 2026-04-30 19:31:11 -07:00
test_hifi_instance_methods.py Fix "no such table: hifi_instances" via defensive lazy-create 2026-05-06 17:49:44 -07:00
test_hls_parsing.py fix: pick OS-specific ffmpeg binary in hls demux fallback test 2026-04-28 21:13:54 -07:00
test_import_album_match_endpoint.py Lift auto-import matching to testable helper + pin contracts 2026-05-09 09:13:23 -07:00
test_import_page_album_lookup_pattern.py Lift auto-import matching to testable helper + pin contracts 2026-05-09 09:13:23 -07:00
test_import_singles_parallel.py Restore soulsync logger state between parallel-imports tests 2026-05-01 08:45:17 -07:00
test_integrity_failure_marks_task_failed.py Fix: tasks showed Completed when file was quarantined 2026-05-03 18:28:32 -07:00
test_library_disk_usage.py Add Library Disk Usage card to System Statistics 2026-05-03 20:17:06 -07:00
test_library_reorganize.py Rewrite Library Reorganize job to delegate to per-album planner 2026-05-06 21:18:20 -07:00
test_library_reorganize_orchestrator.py Drop flaky log-assertion in watchdog test, keep behavioural assertion 2026-05-10 00:01:03 -07:00
test_library_track_identity.py Make extract_external_ids recognize all source-tagging conventions 2026-05-02 18:26:12 -07:00
test_lidarr_download_client.py Fix three Lidarr bugs that prevented it from being a real download source 2026-05-03 15:49:38 -07:00
test_listening_stats_batch_queries.py test: listening stats worker batched query paths 2026-04-19 15:22:25 -07:00
test_manual_pick_no_auto_retry.py Manual picks: stream results, don't auto-retry, fix stuck-at-0% 2026-05-08 15:12:58 -07:00
test_manual_search_endpoint.py Manual picks: stream results, don't auto-retry, fix stuck-at-0% 2026-05-08 15:12:58 -07:00
test_missing_cover_art.py Fix: maintenance findings badge inflated when scan dedup-skipped 2026-05-04 08:55:13 -07:00
test_personalized_playlists_id_gate.py Discover: sharpen track selection (diversity, source-aware popularity, library dedup, SQL genre) 2026-05-08 08:49:22 -07:00
test_provenance_id_persistence.py Persist source IDs at download time + backfill onto tracks on sync 2026-05-02 17:44:10 -07:00
test_qobuz_credential_sync.py Address Copilot review nits 2026-05-05 11:22:01 -07:00
test_reorganize_db_methods.py Reorganize queue: race + dedupe fixes from kettui review 2026-04-26 08:40:24 -07:00
test_reorganize_queue.py Reorganize queue: race + dedupe fixes from kettui review 2026-04-26 08:40:24 -07:00
test_reorganize_runner.py Library reorganize: FIFO queue with live status panel 2026-04-25 18:01:32 -07:00
test_repair_worker_album_fill.py Fix CI test failures from incomplete dummy config and encoding 2026-04-19 13:04:55 -07:00
test_replaygain_summary_parse.py Fix: ReplayGain wrote same +52 dB gain to every track 2026-05-03 19:08:35 -07:00
test_request_cleanup_timer.py test: api/request periodic cleanup timer 2026-04-19 15:22:25 -07:00
test_runtime_state.py Tighten metadata and import safety 2026-04-27 20:28:05 +03:00
test_script_split_integrity.py Unify artist detail: route source artists to standalone page, retire inline Artists page 2026-04-22 17:00:52 -07:00
test_socketio_cors.py Socket.IO CORS: handle self-review nits 2026-04-26 19:24:43 -07:00
test_soundcloud_client.py Lift shared download dataclasses + boot via singleton factory 2026-05-05 09:08:39 -07:00
test_spa_deep_linking.py Rename Search page id from 'downloads' to 'search', bump to 2.43 2026-04-22 13:22:49 -07:00
test_status_engine_fallback.py Manual picks: stream results, don't auto-retry, fix stuck-at-0% 2026-05-08 15:12:58 -07:00
test_sync_discovery_scan_events.py Realign test file names 2026-04-17 21:25:01 +03:00
test_tidal_auth_instructions.py Show Tidal callback port (not Spotify's) in auth instructions 2026-04-30 20:54:06 -07:00
test_tidal_auth_redirect_uri.py Honor configured Tidal redirect_uri, drop request-host fallback 2026-05-02 19:20:18 -07:00
test_tidal_search_shortening.py Reject Tidal streams that silently downgrade from the requested quality 2026-04-24 13:12:30 -07:00
test_tool_progress_events.py Realign test file names 2026-04-17 21:25:01 +03:00
test_track_artist_matching.py Add fallback negative-case test for track-artist matching 2026-04-30 17:13:15 -07:00
test_track_number_repair.py Fix CI test failures from incomplete dummy config and encoding 2026-04-19 13:04:55 -07:00
test_track_search_single_query.py test: api_search_tracks single-query track search 2026-04-19 15:22:24 -07:00
test_unknown_artist_fixer.py Fix test using old deezer_track_id column name 2026-04-20 21:11:45 -07:00
test_watchlist_album_match.py Stop watchlist re-downloading compilation tracks; catch slskd dedup orphans 2026-05-01 12:57:50 -07:00
test_watchlist_bulk_add.py Bulk watchlist add: fall back through every source ID, not just active 2026-04-30 20:27:42 -07:00
test_watchlist_scanner_scan.py Move the import flows and related post-processing pipelines into separate modules 2026-04-27 19:54:43 +03:00
test_websocket_client_transport.py Repoint websocket transport test to core.js after split 2026-04-22 11:30:39 -07:00
test_websocket_infrastructure.py Split metadata source and Spotify status 2026-05-02 22:02:00 +03:00
test_wishlist_pagination.py test: wishlist SQL pagination and category filter 2026-04-19 15:22:24 -07:00
test_worker_existing_id_marks_matched.py test: enrichment worker re-processing fix and migration backfill 2026-04-19 15:22:24 -07:00
test_worker_utils_album_track_count.py Fix Album Completeness job reporting zero findings for every album 2026-04-24 12:39:41 -07:00
test_youtube_ffmpeg_no_eager_download.py ffmpeg gate also covers is_available — fixes the actual leak path 2026-05-08 15:46:31 -07:00