- PreBuildPlaylistItemsCacheAsync was failing with HTTP 401
- Background services don't have client headers for authentication
- Now manually creates X-Emby-Authorization header with API key
- Fixes 'Failed to fetch Jellyfin playlist items: HTTP 401' warning
- Playlist items cache now builds successfully after track matching
All 225 tests pass.
- Fix playlist counting logic to use fuzzy matching (same as track view)
- Count local tracks by matching Jellyfin tracks to Spotify tracks
- Count external matched tracks from cache
- Count missing tracks (not found locally or externally)
- Progress bars now show three colors:
* Green: Local tracks in Jellyfin
* Orange: External matched tracks (SquidWTF/Deezer/Qobuz)
* Grey: Missing tracks (not found anywhere)
- Add 'Missing' badge to tracks that couldn't be found
- Missing tracks can still be manually mapped
- Fixes incorrect counts like '28 matched • 1 missing' showing 29 external tracks
All 225 tests pass.
- Backend now returns mapped track details after saving
- Frontend updates track in-place without requiring page refresh
- Track status changes from External to Local immediately
- Map button is removed after successful mapping
- Playlist counts refresh in background
- Improved UX: no more 'refresh the playlist' message
All 225 tests pass.
- Fix manual mapping track selection visual feedback (use accent color + background)
- Clear all playlist caches after manual mapping (matched, ordered, items)
- Strip [S] suffix from titles/artists/albums when searching for lyrics
- Add [S] suffix to artist and album names when song has [S] for consistency
- Ensures external tracks are clearly marked across all metadata fields
All 225 tests pass.
- Pre-build playlist items cache during track matching for instant serving
- Add PreBuildPlaylistItemsCacheAsync() to SpotifyTrackMatchingService
- Combines local Jellyfin tracks + external matched tracks in correct Spotify order
- Saves to both Redis and file cache for persistence across restarts
- Change matching interval from hardcoded 30 minutes to configurable (default: 24 hours)
- Add SPOTIFY_IMPORT_MATCHING_INTERVAL_HOURS environment variable
- Set to 0 to only run once on startup (manual trigger still works)
- Add endpoint usage files to .gitignore
- Update documentation in README and .env.example
Rationale: Spotify playlists like Discover Weekly update once per week,
so running every 24 hours is more than sufficient. Pre-building the cache
eliminates slow 'on the fly' playlist building.
All 225 tests pass.
- External tracks are injected on-the-fly, not stored in Jellyfin DB
- Check spotify:matched:ordered cache to get accurate external count
- Calculate external tracks as: total matched - local tracks
- This will properly show the two-color progress bar (green local + orange external)
- All 225 tests passing
- Made search query clickable - opens provider-specific search
- Added searchProvider() function with URLs for Deezer, Qobuz, SquidWTF
- Enhanced console logging to debug progress bar data
- Logs raw playlist data and calculated percentages
- Search link opens in new tab
- Increased progress bar height from 8px to 12px for better visibility
- Changed colors to more vibrant shades (green #10b981, orange #f59e0b)
- Added console debug logging for playlist stats
- Shows local (green) and external (orange) track percentages side-by-side
- Fixed external track detection (check for provider prefix in ID)
- Added genre support to MusicBrainz service (inc=genres+tags)
- Created GenreEnrichmentService for async genre lookup with caching
- Show provider name and search query for external tracks in admin UI
- Display search query that will be used for external track streaming
- Aggregate playlist genres from track genres
- All 225 tests passing
- Fixed AdminController export/import .env endpoints (moved from ConfigUpdateRequest class)
- Added ArtistId and AlbumId to integration test fixtures
- All 225 tests passing
- Version set to 1.0.0 (semantic versioning)
- MusicBrainz service ready for future ISRC-based matching (1.1.0)
- Import/export handles full .env configuration with timestamped backups
- Added MusicBrainzSettings model with username/password authentication
- Created MusicBrainzService with ISRC lookup and recording search
- Implements proper rate limiting (1 req/sec) per MusicBrainz rules
- Added meaningful User-Agent header as required
- Registered service in Program.cs with configuration
- Added MusicBrainz section to appsettings.json
- Credentials stored in .env (MUSICBRAINZ_USERNAME/PASSWORD)
Next: Add to admin UI and implement import/export for .env
Verified against real Jellyfin responses for tracks, albums, artists, and playlists:
- Reordered fields to match Jellyfin's exact field order
- Added missing fields: PremiereDate, HasLyrics, Container, ETag, etc.
- Fixed MediaType to 'Unknown' for albums/artists (not null)
- Fixed UserData.Key format to match Jellyfin patterns
- Added ParentLogoItemId, ParentBackdropItemId for proper hierarchy
- Fixed Genres/GenreItems to always be arrays (never null)
- Added complete MediaStream structure with all Jellyfin fields
- Playlists now have MediaType='Audio' to match real playlists
- All responses now perfectly mimic real Jellyfin structure
- Fixed UpdateSpotifyPlaylistCounts to properly handle file cache without skipping items
- Added Genres and GenreItems fields to all tracks (empty array if no genre)
- Added complete MediaStreams with audio codec info for external tracks
- Added missing MediaSource fields: IgnoreDts, IgnoreIndex, GenPtsInput, HasSegments
- Ensured Artists array never contains null values
- All external tracks now have proper genre arrays to match Jellyfin structure
- Added MediaStreams, MediaAttachments, Formats as empty arrays instead of null
- Added RunTimeTicks field to MediaSources
- Added detailed logging to GetExternalChildItems to debug artist album issues
- This should fix 'type Null is not a subtype of type List<dynamic>' error
- Fixed LrclibService.GetLyricsAsync call to use empty string and 0 for duration
- Fixed nullability warnings in SpotifyTrackMatchingService by explicitly casting to nullable tuple
- Blue segment shows local tracks percentage
- Yellow segment shows external matched tracks percentage
- Bar fills to 100% when all tracks are matched
- Added tooltips showing track counts on hover
- Added all required boolean fields to MediaSources (IsRemote, IsInfiniteStream, RequiresOpening, etc)
- UpdateSpotifyPlaylistCounts now loads from file cache if Redis is empty
- This fixes 'type Null is not a subtype of type bool' error in Finamp
- Playlist counts now show correctly even after container restart
- Created ParallelMetadataService to race all providers and return fastest result
- Search now uses parallel service when available for lower latency
- Pre-fetch LRCLib lyrics for top 3 search results in background
- FuzzyMatcher already handles apostrophe normalization (applied everywhere)
- Enhanced FuzzyMatcher to normalize apostrophes (', ', ', etc) for better matching
- Added Redis-only caching for search results (15 min TTL)
- Added pattern-based cache deletion for search and image keys
- Added URL input field in Map to Local modal to paste Jellyfin track URLs
- Added /api/admin/jellyfin/track/{id} endpoint to fetch track details by ID
- Fixed duplicate cache key declaration in GetSpotifyPlaylistTracksOrderedAsync
- Updated cache clearing to include new spotify:playlist:items:* keys
- Properly parse and modify Fields parameter instead of appending to end
- Fixes 400 BadRequest errors from malformed URLs
- Now correctly adds MediaSources to Fields parameter value
CRITICAL FIX: Don't convert Jellyfin items to Song objects and back!
- Get raw Jellyfin playlist items with MediaSources field
- Reorder them according to Spotify positions
- Inject external tracks where needed
- Return raw items preserving ALL Jellyfin metadata (bitrate, etc)
This ensures local tracks in Spotify playlists show correct bitrate
just like regular Jellyfin playlists.
- Avoid JavaScript escaping issues with inline onclick handlers
- Use data attributes to store track info
- Add event listeners after DOM is created
- Prevents syntax errors with special characters in track names
- Ensure MediaSources field is included when proxying browse requests
- Applies to /Users/{userId}/Items (favorites, recently played, etc)
- Jellyfin doesn't include MediaSources by default, must be requested
- Now bitrate info shows correctly in all browse contexts
- Changed ChildCount to reflect tracks actually in Jellyfin (local + external matched)
- Previously was incorrectly adding missing tracks to the count
- Now clients see the correct number of playable tracks
- Uses spotify:matched:ordered cache to count external matched tracks
Backend changes:
- Distinguish between local tracks (in Jellyfin library) and external tracks (downloaded)
- Track external matched vs external missing counts
- Calculate completion percentage for each playlist
Frontend changes:
- Show detailed breakdown: X local • Y matched • Z missing
- Display completion percentage with progress bar
- Color-coded stats (green=local, blue=matched, yellow=missing)
- Updated table headers for clarity
- Added clearing of spotify:matched:* and spotify:matched:ordered:* keys
- This ensures bitrate metadata fix takes effect after cache clear
- Returns count of cleared Redis keys in response
- Add 'Map to Local' button for external tracks in playlist viewer
- Search Jellyfin library to find local tracks
- Save manual mappings (Spotify ID → Jellyfin ID) in cache
- Manual mappings take priority over fuzzy matching
- Clear playlist cache when mapping is saved to force refresh
- UI shows which tracks are manually mapped in logs
CRITICAL FIX: Changed matching strategy completely
- Step 1: Match ALL Jellyfin tracks to Spotify positions (fuzzy 70%)
- Step 2: Build playlist in Spotify order using matched Jellyfin tracks
- Step 3: Fill remaining gaps with external tracks (cached or on-demand)
- Step 4: Add any unmatched Jellyfin tracks at the end
This ensures Jellyfin tracks are ALWAYS used when they match, preventing
external tracks from being used when local versions exist.
- Search external providers in real-time if no cached match exists
- Use fuzzy matching (60% threshold) for external tracks
- Ensures external tracks are always available even without pre-matching
- Local tracks still prioritized first (70% threshold)
- Remove Spotify ID/ISRC matching (Jellyfin plugin doesn't add these)
- Use ONLY fuzzy name matching (title + artist, 70% threshold)
- LOCAL tracks ALWAYS used first before external providers
- Include ALL tracks from Jellyfin playlist (even if not in Spotify)
- Prevent duplicate track usage with HashSet tracking
- AdminController also updated to match by name for Local/External badges
- Better logging with emojis for debugging
- Add fuzzy matching by title+artist as fallback (like Jellyfin Spotify Import plugin)
- Add clear error messages when JELLYFIN_USER_ID is not configured
- Add emoji logging for easier debugging (🔍📌✅❌)
- Check HTTP status code when fetching playlist items
- This should fix the issue where all tracks show [S] even when they exist locally
- CRITICAL FIX: Add UserId parameter to all Jellyfin playlist item fetches (fixes 400 BadRequest errors)
- Fix GetPlaylists to correctly count local/missing tracks
- Fix GetSpotifyPlaylistTracksOrderedAsync to find local tracks (was serving external tracks for everything)
- Fix SpotifyTrackMatchingService to skip tracks already in Jellyfin
- Add detailed debug logging for track matching (LOCAL by ISRC/Spotify ID, EXTERNAL match, NO MATCH)
- Add 'Match Tracks' button for individual playlists (not all playlists)
- Add 'Match All Tracks' button for matching all playlists at once
- Add JELLYFIN_USER_ID to web UI configuration tab for easy setup
- Add /api/admin/playlists/match-all endpoint
This fixes the issue where local tracks weren't being used - the system was downloading
from SquidWTF even when files existed locally in Jellyfin.
