- services/sync_service.py: dropped unused PlexClient / JellyfinClient
/ NavidromeClient class imports. After the engine refactor the
service only needs TrackInfo for type annotations; the class
imports were dead.
- WHATS_NEW: extended the media server engine review-pass entry to
cover the followup commits (Cin-5 per-server global removal +
Gap 1 shared types lift) so the changelog matches the actual
branch state.
// --- post-2.4.1 dev work — entries hidden by _getLatestWhatsNewVersion until the build version bumps ---
{date:'Unreleased — 2.4.2 dev cycle'},
{title:'Internal: Media Server Engine Cin/JohnBaumb Pass',desc:'internal — applied the same architectural cleanups the download engine PR went through to the media server engine PR before review. (1) every server client (Plex / Jellyfin / Navidrome / SoulSync) now explicitly inherits `MediaServerClient` instead of relying on structural typing — drift in any class fails at the conformance test boundary. (2) generic accessors on the engine: `configured_clients()` (replaces per-server `if X and X.is_connected(): clients[name] = X` chains in web_server.py) and `reload_config(name=None)` (generic dispatch instead of per-client reload calls). (3) singleton factory: `get_media_server_engine()` / `set_media_server_engine()` matching the metadata + download engine shape. web_server.py boots via `set_media_server_engine(...)` so factory + global handle share state. (4) ~70 direct `plex_client.X` / `jellyfin_client.X` / `navidrome_client.X` / `soulsync_library_client.X` attribute reaches in web_server.py migrated to `media_server_engine.client(\'<name>\').X`. ~60 standalone refs (truthy checks, media_client assignments, source-name tuples) also routed through the engine. the per-server if/elif chains stay because the work is genuinely server-specific (Plex raw API vs Jellyfin / Navidrome client methods returning different shapes), but the per-server CLIENT REACH now goes through the engine like the POC pattern intended. zero behavior change.' },
{title:'Internal: Media Server Engine Cin/JohnBaumb Pass',desc:'internal — applied the same architectural cleanups the download engine PR went through to the media server engine PR before review. (1) every server client (Plex / Jellyfin / Navidrome / SoulSync) now explicitly inherits `MediaServerClient` instead of relying on structural typing — drift in any class fails at the conformance test boundary. (2) generic accessors on the engine: `configured_clients()` (replaces per-server `if X and X.is_connected(): clients[name] = X` chains in web_server.py) and `reload_config(name=None)` (generic dispatch instead of per-client reload calls). (3) singleton factory: `get_media_server_engine()` / `set_media_server_engine()` matching the metadata + download engine shape. web_server.py boots via `set_media_server_engine(...)` so factory + global handle share state. (4) ~70 direct `plex_client.X` / `jellyfin_client.X` / `navidrome_client.X` / `soulsync_library_client.X` attribute reaches in web_server.py migrated to `media_server_engine.client(\'<name>\').X`. ~60 standalone refs (truthy checks, media_client assignments, source-name tuples) also routed through the engine. (5) the per-server `plex_client` / `jellyfin_client` / `navidrome_client` / `soulsync_library_client` globals in web_server.py are gone entirely — engine owns the client instances now, every caller reaches via `media_server_engine.client(\'<name>\')`. four multi-client consumers (`PlaylistSyncService`, `ListeningStatsWorker`, `WebScanManager`, discovery `SyncDeps`) refactored to take the engine instead of separate per-server kwargs. (6) `TrackInfo` and `PlaylistInfo` lifted out of `core/plex_client.py` / `jellyfin_client.py` / `navidrome_client.py` (each was defining a near-identical copy) into the neutral `core/media_server/types.py` module — same lift Cin caught on the download `TrackResult`/`AlbumResult`/`DownloadStatus` situation. consumers (matching engine, sync service) get one import. zero behavior change.' },
{title:'Internal: Media Server Engine Foundation',desc:'internal — companion to the download engine refactor. introduces a media-server engine + plugin contract on top of the four server clients (plex / jellyfin / navidrome / soulsync standalone). web_server.py historically had ~33 `if active_server == "plex" / "jellyfin" / ...` dispatch sites. new `core/media_server/` package provides `MediaServerEngine` that reads `server.active` config + routes to the right client. plugin contract narrowly requires only the four methods every server actually implements (is_connected, ensure_connection, get_all_artists, get_all_album_ids); optional methods (search_tracks, trigger_library_scan, get_library_stats, etc.) are routed with safe defaults so SoulSync standalone (no library scan API since it walks the filesystem directly) doesn\'t need stub methods. lifted the four uniform `is_connected` dispatches into `engine.is_connected()`. honest scope: most "dispatch sites" the recon counted are genuinely different per server (playlist track replace, per-server metadata sync, deep scan with server-specific cache strategies) — those stay explicit per the "lift what\'s truly shared" standard. 35 new tests pin: per-server observable behavior (4 server pinning files, 21 tests), engine cross-server dispatch (10 tests), structural conformance (4 tests). engine reference + plugin contract available for future targeted refactors. zero behavior change for users.'},
{title:'Internal: Typed Metadata Foundation',desc:'internal — first step of a multi-pr migration to give the metadata pipeline a real contract. the codebase historically grew duck-typed extractors (`_extract_lookup_value(album_data, "id", "album_id", "collectionId", "release_id", default=...)`) at every consumer site because each provider returns its own response shape. ~150 of those across the codebase. new `core/metadata/types.py` defines canonical typed `Album` / `Track` / `Artist` dataclasses with strict required fields. per-source classmethod converters (from_spotify_dict, from_itunes_dict, from_deezer_dict, from_discogs_dict, from_musicbrainz_dict, from_hydrabase_dict) are the SINGLE place that knows each provider\'s wire shape. zero behavior changes in this pr — pure additive foundation. follow-up prs migrate consumers one at a time. full migration plan documented at docs/metadata-types-migration.md.',page:'library'},
{title:'Internal: Migrate Album-Info Builders to Typed Path',desc:'internal — steps 2+3 of the typed metadata migration in one pr. two album-info builders now route through `Album.from_<source>_dict()` when the caller passes a known source: `_build_album_info` (used by every album-tracks lookup) and the embedded album section of `_build_single_import_context_payload` (used by single-track import context resolution). legacy duck-typed extraction stays as the fallback when source is empty/unknown, raw input isn\'t a dict, or the typed converter raises — so a converter bug can\'t break album resolution or import context. caller-provided album_id / album_name / artist_name fallbacks apply on the typed path the same way they did on legacy. zero behavior change for existing callers since they don\'t pass a source yet — opt-in only. 22 new tests pin the typed path, the legacy fallback, and parametrized coverage across registered providers.'},
Broque Thomas
2 weeks ago