Upgrades all references of pgvector/pgvector from pg16 to pg18,
including docker-compose files and e2e testing environments.
Updates the development docker-compose to mount the postgres volume
at /var/lib/postgresql to align with pg18+ directory requirements.
Also adds `shrink-0` to WhatsNewDialog header and footer to resolve
a layout overflow glitch.
Add a next steps checklist to the pairing success card and a permanent help card to the sync settings tab. Explain that books will only sync to the Kobo device if they are placed in a collection with Kobo sync enabled.
Closes#522
Ensure that fixed-layout EPUB content structures expose their frame index and overlayer attributes during retrieval. This prevents the navigation subsystem from falling back to the final page of the document due to unverified page position indexes.
Closes#541
Prevent duplicate annotations from being created when a user modifies the style or color of an existing highlight. Route the popover actions through a PATCH request on the existing annotation ID instead of invoking the creation flow.
Closes#543
When a matched Hardcover edition was missing page counts, progress sync
would silently skip sending updates and save a null progress value to the
local DB. This caused the book to be perpetually stuck in a "Pending sync"
state and triggered endless redundant sync attempts.
We now throw a clear missing pages error during progress calculation to
correctly register the sync error state, log the failure, and trigger background
sync immediately when sync is first enabled for a book.
Closes#544
When transitioning from the series detail view to another view (like edit
metadata), route params update before the component unmounts. This causes
the seriesId computed property to change to null, triggering the seriesId
watcher to reload/reset the books state, which in turn calls setBookContext
with empty values.
Prevent this reset by ignoring null series IDs in both the seriesId watcher
and the books/total watcher, keeping the navigation context intact for the
next view.
Closes#540
When saving metadata synchronously, if the file write operation fails, the rename operation is now skipped to avoid renaming a file that could be left in an inconsistent state.
Closes#532
The dashboard query now joins user book statuses to exclude books marked as 'read', 'skimmed', or 'abandoned' from the continue reading list.
Closes#528
Previously, attempting to rename a book into a folder that already existed on disk was blocked. This change allows users to move books into existing folders, enabling scenarios where different formats (like ebooks and audiobooks) can coexist in the same directory. If a book is already tracked at the target path, the files are merged.
Closes#523
When renaming/organizing a book with multiple audio content files (such as
a multi-track audiobook), only the primary file was renamed to the new
destination path, leaving the other track files to collide or remain
unprocessed.
Retrieve the `sortOrder` for all files in the file-rename repository. When
renaming, detect if a book has multiple audio content tracks and resolve any
target path collisions among them by appending `-PartXX` suffixes, ordered by
their sort position.
Closes#521
Add configuration settings and backend support to define default library access permissions for newly created users. Newly registered users automatically receive these default library access permissions upon creation. Also align current users page controls and magic link options in the admin tab settings UI, adding styled callout cards for empty states.
Closes#486
When manually syncing a book, the plugin uploaded progress without checking if another device had written a newer progress state to the server first, causing reading progress to be overwritten and lost.
Fetch the remote progress from the server before starting a manual sync. If a newer progress state is found on the server, prompt the user to apply the remote progress locally first. If the prompt is rejected, proceed with the sync of annotations and reading history while bypassing the progress upload.
Closes#461
Pass the user's timezone settings to the database query to ensure the
Reading Clock widget groups peak reading hours correctly, instead of
defaulting to UTC.
Also includes a minor UI update to the KOReader plugin to expose a
download options button when multiple formats are available.
Closes#513
Records success and failure events for Kobo device syncs in the database, including operation duration, event types, and error details. Surfaces the recent sync history in the Kobo Settings UI for users to monitor their device activity.
Use request generation counters in composables to ignore in-flight fetches that resolve after a state reset. This ensures that user-scoped data like smart scopes, libraries, and collections do not persist across user switches.
Closes#510
When using date operators in smart scopes or book queries, PostgreSQL
threw an "operator does not exist: date >= integer" error. This explicitly
casts the days offset to an integer in the SQL generated by the book query
builder to prevent the crash.
Closes#508
Some PDF files contain malformed metadata that pdf-lib cannot correctly
decode, despite being readable by external tools. This introduces a
fallback to extract metadata using Poppler's pdfinfo utility, and improves
XMP keyword parsing as a fallback for missing tags.
Closes#491
Processes database queries in chunks of 500 within the migration
repository. This prevents exceeding PostgreSQL maximum bind parameter
limits when importing or querying large datasets.
Closes#504
Introduces the koreader_unmatched_book_devices schema to track unmatched
books per device. Adds API endpoints and UI in settings for dismissing
unmatched books and completely removing a KOReader device.
Closes#504
Keep reader sidebar navigation open for repeated jumps when users pin the panel. Use session-scoped state so each book file can remember the preference.
Closes#497
When KOReader syncs progress or annotations, it identifies books by a document
hash. BookOrbit normally resolves this automatically, but some books cannot be
matched. This adds a tracking table for unmatched hashes and a UI to manually
link them to existing library books.
Closes#458
During an automatic metadata refresh, the fetched Hardcover edition ID was not being copied to the update payload. This ensures the value is persisted to the database alongside the other metadata fields.
Closes#440
Adds a fixedLayoutSpread preference to EPUB reader settings to allow forcing single-page display. This is particularly useful for image-based EPUBs like comics and manga.
Closes#487
Add support for selecting multiple books and downloading entire lists, pages, or selections in bulk. This includes long-press actions, a dedicated bulk download queue with retry and cancel support, and format preset configurations.
The commit also simplifies the OPDS random book query using ORDER BY random() and adjusts the Nest dev CLI config to disable the Swagger plugin for faster restarts.
Closes#478
Replace the plain textarea for book descriptions with a Tiptap-based rich text editor in the metadata edit tab. Include HTML sanitization utilities to ensure safe rendering of user-provided descriptions.
Closes#460
Reading source is often less relevant for day-to-day analytics than
book format or reading pace. Shift the source distribution chart to the
end of the user dashboard and change the default breakdown dimension
from source to format to prioritize more useful metrics. Include a
migration to update existing users who have not customized their chart
order.
Progress values computed during reader events could occasionally result in NaN, Infinity, or out-of-bounds numbers, which caused API validation failures or corrupted state. Progress values are now clamped to valid ranges and sanitized to omit non-finite numbers before tracking.
Closes#479
The "Discover Something New" dashboard section previously only excluded
books based on reading progress. It now joins the userBookStatus table
to correctly exclude books that are actively being read, finished, or
abandoned.
Closes#484
Import Booklore and Grimmory reading sessions, personal ratings, and progress state through the migration pipeline.
Preserve book-level progress as primary-file progress when source file progress rows are absent.
Closes#482Closes#483
The search dropdown previously limited results to a fixed number. This change implements virtual scrolling and lazy loading in the AppHeader search bar to allow users to scroll through an unlimited number of matching books efficiently.
Closes#480
The group rule validator now accepts an optional `provider` field for
`communityRating` rules. This enables smart scopes to filter books based
on community ratings from specific providers (e.g., Goodreads, Amazon)
rather than just the aggregate rating.
The validator ensures that the provider field is only applied to community
rating rules and rejects unsupported providers.
Closes#477
* feat(metadata): parse Calibre #extra_tags custom column into book tags
Extract #extra_tags from Calibre user metadata and populate ParsedOpf.tags when
the custom column is present.
Support JSON array, pipe-delimited, and comma-delimited string values. Merge
tags from #extra_tags with bookorbit:tags and deduplicate them.
* fix(opf-parser): prefer pipe over comma when splitting extra tags string
Previously, coerceExtraTags used the regex /[|,]/ which splits on both
delimiters simultaneously. A value like "History, Modern|Fantasy" was
parsed into three tags instead of two because every comma and pipe was
treated as a separator regardless of which was the primary delimiter.
Fix: check if the string contains a pipe first; if so, split only on "|",
otherwise fall back to splitting on ",". This preserves commas embedded
within pipe-delimited tags (e.g. "History, Modern" stays intact).
Also adds a regression test covering the mixed-delimiter case.
Adds support for displaying custom metadata as columns in the book
table view. The frontend now fetches active fields and dynamically
renders the corresponding cells, backed by book card API updates.
Closes#465
ISBN queries were sent to ItemSearch.aspx as keywords, missing ~24%
of valid Korean ISBNs. Route ISBN searches through ItemLookUp.aspx
with ItemIdType=ISBN13 (13-digit) or ISBN (10-digit), falling back
to keyword search for unrecognised formats.
Closes#437
Co-authored-by: neon <18559658+neonsolstice@users.noreply.github.com>
Previously, the explicit metadata search modal for a book was restricted
to only the providers configured in the library's field rules. This
prevented users from manually searching a specific provider if it
was not part of their default automated fetch pipeline.
This change allows all globally enabled providers to be used during
a manual metadata search, while still defaulting the selected UI
checkboxes to the providers specified in the field rules.
The KOReader plugin previously matched series by name, which caused
issues for series with identical names or special characters. This
updates the plugin and backend to use the exact series ID for
navigation and filtering, ensuring all series load reliably. Also
bumps the KOReader plugin version to 1.0.1.
Closes#466
Extracts the ZIP index parser into a dedicated utility that avoids
loading the entire file into a single memory buffer. This fixes an
issue where CBZ files larger than 2GB would fail to load due to
Node.js buffer limits. Also extracts CBR error handling into helper
functions for readability.
Closes#457
Introduces Swagger and OpenAPI documentation generation for the
backend API. Configures bearer authentication globally while
allowing explicit bypasses for public endpoints. Exports nested
DTO classes across several modules so the NestJS Swagger plugin
can correctly reflect their schema definitions.
Closes#399
Routes multiple book selection to the edit tab instead of details.
Fixes a bug where deactivated kept-alive sources would passively
clobber the active navigation context by tracking component
activation state. Converts route param computed refs to shallow
refs to prevent reactivity issues during navigation.
Closes#455
The drizzle-orm query builder hits a RangeError when generating a single
bulk INSERT for tens of thousands of authors during backfill. Chunking
the insertions into smaller batches resolves the issue.
Closes#452
Introduce BookDockWorkQueue to serialize metadata extraction and auto-finalize
jobs. Files process newest-first with a 250 ms drain window and 500-1000 ms
inter-job spacing to avoid hammering metadata providers. Watched-folder rescans
now requeue files stuck in extracting or fetching status after a server restart
rather than silently dropping them.
Replace the global ThrottlerGuard with SensitiveEndpointThrottlerGuard, which
only activates on endpoints that carry explicit @Throttle() metadata. This
eliminates all @SkipThrottle() decorators from non-auth controllers and adds
missing throttle coverage for oidcGenerateLinkState and oidcGeneratePreviewState.
Update the login page to match the simplified 429 error message.
Fix countsByStatus to aggregate extracting and fetching rows into the pending
bucket so the UI counter stays accurate mid-pipeline. Add a secondary id sort
to findAll for deterministic pagination ordering.
Auto-detect the container memory limit via cgroup v1/v2 in entrypoint.sh and
compute --max-old-space-size proportionally (50-75 % of limit) instead of
defaulting to a fixed 2048 MB that OOMs small containers.
Closes#441
Pass the authenticated API fetcher to the streaming book loader instead of a static token. This ensures the token is automatically refreshed if it expires while reading, preventing 401 Unauthorized errors when navigating between chapters.
Closes#446
neon
dependabot[bot]
Jadehawk
Kyungwook Park