Add multi-source ebook search & processing

Refactor ebook flow to support multiple sources (Anna's Archive direct downloads + Prowlarr indexer search) and unify handling with existing audiobook processors. Key changes:
- search-ebook.processor: rewritten to try Anna's Archive first then fall back to indexer search, add Prowlarr grouping, ranking (rankEbookTorrents), and handlers to route results to direct-download or download-torrent flows.
- organize-files.processor: enriches audiobook/ebook metadata from AudibleCache (year, narrator), treats indexer downloads specially (seed retention), adds optional NZB cleanup/archive logic, and improves retryable error detection.
- file-organizer: organizeEbook now accepts additional metadata and an isIndexerDownload flag and supports directories vs single-file paths.
- API/UI: include request.type in admin requests API and remove the “coming soon” notice from Ebook settings tab.
- fetch-ebook route: removed blocking error for indexer-only mode so the flow can proceed when indexer search is enabled.
- Documentation: update TOC, ebook-sidecar, settings-pages, and ranking-algorithm docs to describe indexer search, unified ebook ranking, configuration, and flows.
These changes enable indexer-based ebook discovery, ranking, and downloads while preserving existing Anna's Archive behavior and reusing audiobook download processors where possible.

documentation/phase3/ranking-algorithm.md

@@ -286,6 +286,80 @@ const ranked = rankTorrents(torrents, audiobook, {
 return ranked;  // User can see torrents without author info
 ```
 
+## Ebook Torrent Ranking
+
+The ranking algorithm also supports ebook torrents from indexers with ebook-specific scoring.
+
+### Unified Code Architecture
+
+Ebook ranking **reuses** the following from audiobook ranking:
+- `scoreMatch()` - Title/author matching (60 pts)
+- `scoreSeeders()` - Seeder count scoring (15 pts)
+- Bonus modifier system (indexer priority, flag bonuses)
+- Dual threshold filtering (base >= 50, final >= 50)
+
+### Ebook-Specific Scoring
+
+**Format Match (10 pts max)**
+- 10 pts if torrent format matches preferred format
+- 0 pts otherwise (no partial credit)
+- Format detected from torrent title keywords: `.epub`, `.pdf`, `.mobi`, `.azw3`, etc.
+
+**Size Quality (15 pts max, INVERTED)**
+- < 5 MB: 15 pts (optimal for ebooks)
+- 5-15 MB: 10 pts (may have images)
+- 15-20 MB: 5 pts (large but acceptable)
+- > 20 MB: **Filtered out** (too large for ebooks)
+
+### Ebook vs Audiobook Comparison
+
+| Component | Audiobook | Ebook |
+|-----------|-----------|-------|
+| Title/Author | 60 pts (reused) | 60 pts (reused) |
+| Format | 10 pts (M4B > M4A > MP3) | 10 pts (match = 10, else 0) |
+| Size | 15 pts (larger = better) | 15 pts (smaller = better) |
+| Seeders | 15 pts (reused) | 15 pts (reused) |
+| Size Filter | < 20 MB filtered | > 20 MB filtered |
+
+### Ebook Interface
+
+```typescript
+interface EbookTorrentRequest {
+  title: string;
+  author: string;
+  preferredFormat: string;  // 'epub', 'pdf', 'mobi', etc.
+}
+
+interface RankEbookTorrentsOptions {
+  indexerPriorities?: Map<number, number>;
+  flagConfigs?: IndexerFlagConfig[];
+  requireAuthor?: boolean;  // Default: true
+}
+
+function rankEbookTorrents(
+  torrents: TorrentResult[],
+  ebook: EbookTorrentRequest,
+  options?: RankEbookTorrentsOptions
+): RankedEbookTorrent[];
+```
+
+### Ebook Usage Example
+
+```typescript
+// Ebook search from indexers
+const ranked = rankEbookTorrents(prowlarrResults, {
+  title: 'Project Hail Mary',
+  author: 'Andy Weir',
+  preferredFormat: 'epub',
+}, {
+  indexerPriorities,
+  flagConfigs,
+  requireAuthor: true,
+});
+
+const bestEbook = ranked[0];  // Safe to auto-download
+```
+
 ## Tech Stack
 
 - string-similarity (fuzzy matching)