- Fix POST requests (series/list, books/list) not being cached by Next.js fetch cache
by wrapping them with unstable_cache in the private fetch method
- Wrap getHomeData() entirely with unstable_cache so all 5 home requests are cached
as a single unit, reducing cold-start cost from 5 parallel calls to 0 on cache hit
- Remove N+1 book count enrichment from getLibraries() (8 extra calls per cold start)
as LibraryDto does not return booksCount and the value was only used in BackgroundSettings
- Simplify getLibraryById() to reuse cached getLibraries() data instead of making
separate HTTP calls (saves 2 calls per library page load)
- Fix cache debug logs: replace misleading x-nextjs-cache header check (always UNKNOWN
on external APIs) with pre-request logs showing the configured cache strategy
- Remove book count display from BackgroundSettings as it is no longer fetched
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Introduced a new resolver function to streamline fetching Stripstream configuration from the database or environment variables.
- Updated various components and API routes to utilize the new configuration resolver, improving code maintainability and reducing direct database calls.
- Added optional environment variables for Stripstream URL and token in the .env.example file.
- Refactored image loading logic in the reader components to improve performance and error handling.
- Introduce provider abstraction layer (IMediaProvider, KomgaProvider, StripstreamProvider)
- Add Stripstream Librarian as second media provider with full feature parity
- Migrate all pages and components from direct Komga services to provider factory
- Remove dead service code (BaseApiService, HomeService, LibraryService, SearchService, TestService)
- Fix library/series page-based pagination for both providers (Komga 0-indexed, Stripstream 1-indexed)
- Fix unread filter and search on library page for both providers
- Fix read progress display for Stripstream (reading_status mapping)
- Fix series read status (books_read_count) for Stripstream
- Add global search with series results for Stripstream (series_hits from Meilisearch)
- Fix thumbnail proxy to return 404 gracefully instead of JSON on upstream error
- Replace duration-based cache debug detection with x-nextjs-cache header
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add LIBRARY_SERIES_CACHE_TAG to getLibrarySeries fetch
- Revalidate library-series tag in updateReadProgress and deleteReadProgress
- Add eslint ignores for temp/, .next/, node_modules/
Made-with: Cursor
Move library header/covers to deterministic server-side rendering, split preference controls into controlled/uncontrolled modes, and remove client cover wrapper to eliminate hydration mismatches and provider coupling on library pages.
- Add cache tags support to BaseApiService
- Tag home data with 'home-data' tag in HomeService
- Use revalidateTag('home-data', 'max') after read progress updates
- With 'max' profile: serve stale while fetching fresh in background
