Use Telegram sendPhoto API for conversion and metadata-approved events
when a book thumbnail is available on disk. Falls back to text message
if photo upload fails.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add notifications crate shared between API and indexer to send Telegram
messages on scan/thumbnail/conversion completion/failure, metadata linking,
batch and refresh events. Configurable via a new Notifications tab in the
backoffice settings with per-event toggle switches grouped by category.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Authors were only sourced from books.authors/books.author fields which are
often empty. Now also aggregates authors from series_metadata.authors
(populated by metadata providers like bedetheque). Adds author filter to
/series endpoint and updates the author detail page to use it.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Extract series code from books.rs into dedicated series.rs module
- Reorganize OpenAPI tags: split overloaded "books" tag into books, series, search, stats
- Add missing endpoints to OpenAPI: metadata_batch, metadata_refresh, komga, update_metadata_provider
- Add missing schemas: MissingVolumeInput, Komga/Batch/Refresh DTOs
- Fix access control: move GET /libraries and POST /libraries/:id/scan to read routes
so non-admin tokens can list libraries and trigger scans
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
When no specific library is selected, iterate over all libraries and
trigger a job for each one, skipping libraries with metadata disabled.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Server and client could use different default locales for
toLocaleDateString/toLocaleString, causing React hydration errors.
Pass the user locale explicitly in JobsList and SettingsPage.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Replace fetch polling in JobsIndicator with EventSource connected to
/api/jobs/stream. Fix the SSE route to return all jobs (via
/index/status) instead of only active ones, since JobsList also
consumes this stream for the full job history. JobsIndicator now
filters active jobs client-side. SSE server-side uses adaptive
interval (2s active, 15s idle) and only sends when data changes.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Make apiFetch support Next.js revalidate option instead of
hardcoding cache: no-store on every request. Stable endpoints
(libraries, settings, stats, series statuses) now use time-based
revalidation while dynamic data (books, search, jobs) stays uncached.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Prevent unnecessary re-renders of book grid items when parent
components update without changing book data.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Dynamic import EditBookForm, EditSeriesForm, MetadataSearchModal, and
ProwlarrSearchModal so their code is split into separate chunks and
only fetched when the user interacts with them.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Pause polling when the tab is hidden, refetch immediately when it
becomes visible again, and use a 30s interval when no jobs are active
instead of polling every 2s unconditionally.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Remove `unoptimized` flag from all thumbnail/cover Image components
and add proper responsive `sizes` props. Convert raw `<img>` tags on
the libraries page to next/image. Add 24h minimumCacheTTL for
optimized images. BookPreview keeps `unoptimized` since the API
already returns optimized WebP.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Eliminates N+1 sequential fetchSeries calls on the libraries page by
returning series count and up to 5 thumbnail book IDs (one per series)
directly from GET /libraries.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Save/restore filter values in LiveSearchForm using localStorage keyed
by basePath (e.g. filters:/books, filters:/series). Filters are restored
on mount when the URL has no active filters, and cleared when the user
clicks the Clear button.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add two new filters to the books listing page:
- Format filter (CBZ/CBR/PDF/EPUB) using existing API support
- Metadata linked/unlinked filter with new API support via
LEFT JOIN on external_metadata_links (using DISTINCT ON CTE
matching the series endpoint pattern)
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Replace thumbnail mosaic with fan/arc layout using series covers as background
- Move library settings from dropdown to full-page portal modal with sections
- Move FolderPicker modal to portal for proper z-index stacking
- Add descriptions to each setting for better clarity
- Move delete button to card header, compact config tags
- Add i18n keys for new labels and descriptions (en/fr)
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add "Deep rescan" job type that clears directory mtimes to force
re-walking all directories, discovering newly supported formats (e.g.
EPUB) without deleting existing data or metadata.
Also improve full rebuild button: red destructive styling instead of
warning, and FR description explicitly mentions metadata/reading status
loss. Rename FR rebuild label to "Mise à jour".
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Parse EPUB structure (container.xml → OPF → spine → XHTML) to extract
images in reading order. Zero new dependencies — reuses zip + regex
crates with pre-compiled regexes and per-file index cache for
performance. Falls back to CBZ-style image listing when spine contains
no images. Includes DB migration, API/indexer/backoffice updates.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The job detail page was only server-rendered with no live updates,
unlike the jobs list page. Add a lightweight JobDetailLive client
component that subscribes to the existing SSE endpoint and calls
router.refresh() on each update, keeping the page in sync while
a job is running.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
API extracts volume numbers from release titles and matches them against
missing volumes sent by the frontend. Matched results are highlighted in
green with badges indicating which missing volumes were found.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add a new GET /authors endpoint that aggregates unique authors from books
with book/series counts, pagination and search. Add author filter to
GET /books. Backoffice gets a list page with search/sort and a detail
page showing the author's series and books.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add metadata_refresh_mode (manual/hourly/daily/weekly) to libraries,
with automatic scheduling via the indexer. Includes API support,
backoffice UI controls, i18n translations, and DB migration.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The Next.js monitoring route was dropping metadata_refresh_mode from the
request body, so the value was never forwarded to the Rust API and
reverted on reload.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Add text input for custom search queries in Prowlarr modal
- Quick search badges pre-fill the input and trigger search
- Default query uses quoted series name for exact match
- Add custom_query support to backend API
- Limit badge area height with vertical scroll
- Add debug logging for Prowlarr API responses
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Move Prowlarr search button before Metadata button
- Add amber collection progress bar showing owned/expected books ratio
- Remove yellow missing count badge from MetadataSearchModal (now shown in progress bar)
- Fix i18n plural parameter for series read count
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Send Prowlarr search results directly to qBittorrent from the modal.
Backend authenticates via SID cookie (login + add torrent endpoints).
- Backend: qbittorrent module with add and test endpoints
- Migration: add qbittorrent settings (url, username, password)
- Settings UI: qBittorrent config card with test connection
- ProwlarrSearchModal: send-to-qBittorrent button per result row
with spinner/checkmark state progression
- Button only shown when qBittorrent is configured
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add Prowlarr indexer integration (step 1: config + manual search).
Allows searching for comics/ebooks releases on Prowlarr indexers
directly from the series detail page, with download links and
per-volume search for missing books.
- Backend: new prowlarr module with search and test endpoints
- Migration: add prowlarr settings (url, api_key, categories)
- Settings UI: Prowlarr config card with test connection button
- ProwlarrSearchModal: auto-search on open, missing volumes shortcuts
- Fix series.readCount i18n plural parameter on series pages
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Make mapped_status nullable so unmapping (X button) sets NULL instead of
deleting the row — provider statuses never disappear from the UI
- normalize_series_status now returns the raw provider status (lowercased)
when no mapping exists, so all statuses are stored in series_metadata
- Fix series_statuses query crash caused by NULL mapped_status values
- Fix metadata batch/refresh server actions crashing page on 400 errors
- StatusMappingDto.mapped_status is now string | null in the backoffice
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
