feat: table series avec UUID PK — migration complète backend + frontend

Migration DB (0070 + 0071):
- Backup automatique de book_reading_progress avant migration
- Crée table series (fusion de series_metadata) avec UUID PK
- Ajoute series_id FK à books, external_metadata_links, anilist_series_links,
  available_downloads, download_detection_results
- Supprime les colonnes TEXT legacy et la table series_metadata

Backend API + Indexer:
- Toutes les queries SQL migrées vers series_id FK + JOIN series
- Routes /series/:name → /series/:series_id (UUID)
- Nouvel endpoint GET /series/by-name/:name pour lookup par nom
- match_title_volumes() factorisé entre prowlarr.rs et download_detection.rs
- Fix scheduler.rs: settings → app_settings
- OpenAPI mis à jour avec les nouveaux endpoints

Frontend:
- Routes /libraries/[id]/series/[name] → /series/[seriesId]
- Tous les composants (Edit, Delete, MarkRead, Prowlarr, Metadata,
  ReadingStatus) utilisent seriesId
- compressVolumes() pour afficher T1→3 au lieu de T1 T2 T3
- Titre release en entier (plus de truncate) dans available downloads

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

apps/indexer/src/job.rs

@@ -17,14 +17,12 @@ async fn rematch_unlinked_books(pool: &PgPool, library_id: Uuid) {
                 b.id AS book_id
             FROM external_book_metadata ebm2
             JOIN external_metadata_links eml ON eml.id = ebm2.link_id
-            JOIN books b ON b.library_id = eml.library_id
+            JOIN books b ON b.series_id = eml.series_id
                 AND b.volume = ebm2.volume_number
-            LEFT JOIN series s ON s.id = b.series_id
             WHERE eml.library_id = $1
               AND ebm2.book_id IS NULL
               AND ebm2.volume_number IS NOT NULL
               AND eml.status = 'approved'
-              AND LOWER(COALESCE(s.name, 'unclassified')) = LOWER(eml.series_name)
         ) matched
         WHERE ebm.id = matched.ebm_id
         "#,

apps/indexer/src/scheduler.rs

@@ -122,7 +122,7 @@ pub async fn check_and_schedule_reading_status_push(pool: &PgPool) -> Result<()>
 pub async fn check_and_schedule_download_detection(pool: &PgPool) -> Result<()> {
     // Only schedule if Prowlarr is configured
     let prowlarr_configured: bool = sqlx::query_scalar(
-        "SELECT EXISTS(SELECT 1 FROM settings WHERE key = 'prowlarr' AND value->>'base_url' IS NOT NULL AND value->>'base_url' != '')"
+        "SELECT EXISTS(SELECT 1 FROM app_settings WHERE key = 'prowlarr' AND value->>'base_url' IS NOT NULL AND value->>'base_url' != '')"
     )
     .fetch_one(pool)
     .await