mirror of
https://github.com/Kareadita/Kavita.git
synced 2025-05-31 12:14:44 -04:00
13 Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
Joseph Milazzo
|
f701f8e599
|
Book Reader Bugfixes (#1254)
* Fixed image scoping breaking and books not being able to load images * Cleaned up a lot of variables and added more jsdoc. After shifting the margin, we try to recover the column layout width,height, and scroll posisiton. * Tap to paginate is working on first load now * On resize, rescroll to attempt to avoid breakage * Fixed transparent background for action bar on white theme * Moved some lists to immutable arrays * Actually fixed backgournd now * Fixed some settings not updating in book reader on load * Put some code in place to test out opening menu with clicking on the document * Fixed settings not propagating to the reader * Fixing 2 column when loading annd ios mobile * Fixed an issue where paging to prev page would sometimes skip the first page. * Fixing previous page skipping first page of chapter * removing console logs * Save progress when we page * Click on document to show the side nav * Removed columns auto because it could render more columns than applicable. Don't explicitly call saveProgress on prev page, as we already do in another call. Adjusted the logic to calculate windowHeight and width to be the same throughout the reader. * Setting select fix and settings polish * Fixed awkward tooltip wording * Added a message for when there is nothing to show on recommended tab * Removed bug marker, there was no bug after all * Fixing book title truncation in action bar * When counting volumes or chapters that have range, count the max part of the range for publication status. * Fixing TOC rendering issue * Styling fixes - Fixed an issue where the image height in the book reader was the column height plus padding so it was breaking pagination calc. - Centered book reader setting pills - Made inactive setting pill into a ghost button - Fixed spacing across the reader settings drawer * Added a bit of code to allow us to disable buttons before we click for next chapter load * Removed titles from action bars * The next page button will now show as the primary color to indicate to the user what the next forward page is. * Added a view series to bookmark page and removed actions from header since it didn't work * Fixed a bug where pagination wasn't mutating url state * Lots of changes, code is kinda working. Added Immersive Mode, but didn't generate migration. Added concept of virtual pages with ability to see them. Math is still slightly off. Cleaned up prefetching code so we do it much earlier. Added some code that doesn't work to disable buttons with virtual paging included. * When turning immersive mode on, force tap to paginate * Refactored out the book reader state as it wasn't very beneficial * Fixed total virtual page calculation * Next/prev page seems to be working pretty well * Applied Robbie's virtual page logic and fixed a bug in prev page code * Changed the next page to use same virtual page logic * Getting back and forward working...somehow. * removing redundant code * Fixing book title overflow from new action bar changes * Polishing pagination styles * Changing chapter to section * Fixing up other book reader themes * Fixed the login header being off-center * Fixing styling to follow approach * Refactored the pagination buttons to properly call next/prev page based on reading direction * Drawer pagination buttons now respect when there is no chapters (prev/next) * Everything except disabling buttons when on last possible page working * Added Book Reader immersive mode migration * Disable next/prev buttons for continuous reading before we request next/prev chapter if there is no chapter. * Show a tooltip for the title * Fixed unit test Co-authored-by: Robbie Davis <robbie@therobbiedavis.com> |
||
|
Joseph Milazzo
|
553f9b0d98
|
Fixes, Tweaks, and Series Filtering (#1217)
* From previous fix, added the other locking conditions on the update series metadata. * Fixed a bug where custom series, collection tag, and reading list covers weren't being removed on cleanup. * Ensure reading list detail has a margin to align to the standard * Refactored some event stuff to use dedicated consts. Introduced a new event when users read something, which can update progress bars on cards. * Added recomended and library tags to the library detail page. This will eventually offer more custom analytics * Cleanup some code onc arousel * Adjusted scale to height/width css to better fit * Small css tweaks to better center images in the manga reader in both axis. This takes care of double page rendering as well. * When a special has a Title set in the metadata, on series detail page, show that on the card rather than filename. * Fixed a bug where when paging in manga reader, the scroll to top wasn't working due to changing where scrolling is done * More css goodness for rendering images in manga reader * Fixed a bug where clearing a typeahead externally wouldn't clear the x button * Fixed a bug where filering then using keyboard would select wrong option * Added a new sorting field for Last Chapter Added (new field) to get a similar on deck feel. * Tweaked recently updated to hit the NFR of 500ms (300ms fresh start) and still give a much better experience. * Refactored On deck to now go to all series and also sort by last updated. Recently Added Series now loads all series with sort by created. * Some tweaks on css for cover image chooser * Fixed a bug in pagination control where multiple pagination events could trigger on load and thus multiple requests for data on parent controller. * Updated edit series modal to show when the last chapter was added and when user last read it. * Implemented a highlight on the fitler button when a filter is active. * Refactored metadata filter screens to perserve the filters in the url and thus when navigating back and forth, it will retain. users should click side nav to reset the state. * Hide middle section on companion bar on phones * Cleaned up some prefilters and console.logs * Don't open drawer by default when a filter is active |
||
|
Joseph Milazzo
|
19678383b2
|
Tech Debt + Series Sort bugfix (#1192)
* Code cleanup. When copying files, if the target file already exists, append (1), (2), etc onto the file (this is enhancing existing implementation to allow multiple numbers) * Added a ton of null checks to UpdateSeriesMetadata and made the code work on the rare case (not really possible) that SeriesMetadata doesn't exist. * Updated Genre code to use strings to ensure a better, more fault tolerant update experience. * More cleanup on the codebase * Fixed a bug where Series SortName was getting emptied on file scan * Fixed a bad copy * Fixed unit tests |
||
|
Joseph Milazzo
|
0e2fafe396
|
Took care of SonarCloud code issues (#1041) | ||
|
Joseph Milazzo
|
0be0e294aa
|
Metadata Optimizations (#910)
* Added a tooltip to inform user that format and collection filter selections do not only show for the selected library. * Refactored a lot of code around when we update chapter cover images. Applied an optimization for when we re-calculate volume/series covers, such that it only occurs when the first chapter's image updates. * Updated code to ensure only lastmodified gets refreshed in metadata since it always follows a scan * Optimized how metadata is populated on the series. Instead of re-reading the comicInfos, instead I read the data from the underlying chapter entities. This reduces N additional reads AND enables the ability in the future to show/edit chapter level metadata. * Spelling mistake * Fixed a concurency issue by not selecting Genres from DB. Added a test for long paths. * Fixed a bug in filter where collection tag wasn't populating on load * Cleaned up the logic for changelog to better compare against the installed verison. For nightly users, show the last stable as installed. * Removed some demo code * SplitQuery to allow loading tags much faster for series metadata load. |
||
|
Joseph Milazzo
|
94bad97511
|
More Filtering and Support for ComicInfo v2.1 (draft) Tags (#851)
* Added a reoccuring task to cleanup db entries that might be abandoned. On library page, the Library in question will be prepoulated. * Laid out the foundation for customized sorting. Added all series page to the UI when clicking on Libraries section header on home page so user can apply any filtering they like. * When filtering, the current library filter will now automatically filter out the options for people and genres. * Implemented Sorting controls * Clear now clears sorting and read progress. Sorting is disabled on deck and recently added. * Fixed an issue where all-series page couldn't click to open series * Don't let the user unselect the last read progress. Added new comicinfo v2.1 draft tags. * Hooked in Translator tag into backend and UI. * Fixed an issue where you could open multiple typeaheads at the same time * Integrated Translator and Tags ComicInfo extension fields. Started work on a badge expander. * Reworked a bit more on badge expander. Added the UI code for Age Rating and Tags * Integrated backend for Tags, Translator, and Age Rating * Metadata tags now collapse if more than 4 present * Some code cleanup * Made the not read badge slightly smaller |
||
|
Joseph Milazzo
|
bbe8f800f6
|
.NET 6 Coding Patterns + Unit Tests (#823)
* Refactored all files to have Interfaces within the same file. Started moving over to file-scoped namespaces. * Refactored common methods for getting underlying file's cover, pages, and extracting into 1 interface. * More refactoring around removing dependence on explicit filetype testing for getting information. * Code is buildable, tests are broken. Huge refactor (not completed) which makes most of DirectoryService testable with a mock filesystem (and thus the services that utilize it). * Finished porting DirectoryService to use mocked filesystem implementation. * Added a null check * Added a null check * Finished all unit tests for DirectoryService. * Some misc cleanup on the code * Fixed up some bugs from refactoring scan loop. * Implemented CleanupService testing and refactored more of DirectoryService to be non-static. Fixed a bug where cover file cleanup wasn't properly finding files due to a regex bug. * Fixed an issue in CleanupBackup() where we weren't properly selecting database files older than 30 days. Finished CleanupService Tests. * Refactored Flatten and RemoveNonImages to directory service to allow CacheService to be testable. * Finally have CacheService tested. Rewrote GetCachedPagePath() to be much more straightforward & performant. * Updated DefaultParserTests.cs to contain all existing tests and follow new test layout format. * All tests fixed up |
||
|
Joseph Milazzo
|
e7619e6b0a
|
Local Metadata Integration Part 1 (#817)
* Started with some basic plumbing with comic info parsing updating Series/Volume. * We can now get chapter title from comicInfo.xml * Hooked in the ability to store people into the chapter metadata. * Removed no longer used imports, fixed up some foreign key constraints on deleting series with person linked. * Refactored Summary out of the UI for Series into SeriesMetadata. Updated application to .net 6. There is a bug in metadata code for updating. * Removed the parallel.ForEach with a normal foreach which lets us use async. For I/O heavy code, shouldn't change much. * Refactored scan code to only check extensions with comic info, fixed a bug on scan events not using correct method name, removed summary field (still buggy) * Fixed a bug where on cancelling a metadata request in modal, underlying button would get stuck in a disabled state. * Changed how metadata selects the first volume to read summary info from. It will now select the first non-special volume rather than Volume 1. * More debugging and found more bugs to fix * Redid all the migrations as one single one. Fixed a bug with GetChapterInfo returning null when ChapterMetadata didn't exist for that Chapter. Fixed an issue with mapper failing on GetChapterMetadata. Started work on adding people and a design for people. * Fixed a bug where checking if file modified now takes into account if file has been processed at least once. Introduced a bug in saving people to series. * Just made code compilable again * Fixed up code. Now people for series and chapters add correctly without any db issues. * Things are working, but I'm not happy with how the management of Person is. I need to take into account that 1 person needs to map to an image and role is arbitrary. * Started adding UI code to showcase chapter metadata * Updated workflow to be .NET 6 * WIP of updating card detail to show the information more clearly and without so many if statements * Removed ChatperMetadata and store on the Chapter itself. Much easier to use and less joins. * Implemented Genre on SeriesMetadata level * Genres and People are now removed from Series level if they are no longer on comicInfo * PeopleHelper is done with unit tests. Everything is working. * Unit tests in place for Genre Helper * Starting on CacheHelper * Finished tests for ShouldUpdateCoverImage. Fixed and added tests in ArchiveService/ScannerService. * CacheHelper is fully tested * Some DI cleanup * Scanner Service now calls GetComicInfo for books. Added ability to update Series Sort name from metadata files (mainly epub as comicinfo doesn't have a field) * Forgot to move a line of code * SortName now populates from metadata (epub only, ComicInfo has no tags) * Cards now show the chapter title name if it's set on hover, else will default back to title. * Fixed a major issue with how MangaFiles were being updated with LastModified, which messed up our logic for avoiding refreshes. * Woohoo, more tests and some refactors to be able to test more services wtih mock filesystem. Fixed an issue where SortName was getting set as first chapter, but the Series was in a group. * Refactored the MangaFile creation code into the DbFactory where we also setup the first LastModified update. * Has file changed bug is now finally fixed * Remove dead genres, refactor genre to use title instead of name. * Refactored out a directory from ShouldUpdateCoverImage() to keep the code clean * Unit tests for ComicInfo on BookService. * Refactored series detail into it's own component * Series-detail now received refresh metadata events to refresh what's on screen * Removed references to Artist on PersonRole as it has no metadata mapping * Security audit * Fixed a benchmark * Updated JWT Token generator to use new methods in .NET 6 * Updated all the docker and build commands to use net6.0 * Commented out sonar scan since it's not setup for net6.0 yet. |
||
|
Joseph Milazzo
|
68bb5ed5a8
|
Feature/bookmark feedback (#508)
* ImageService had a stream reset before writting out to array. Added logging statment for updating series metadata. Removed ConcurencyCheck due to bad update issue for CollectionTag. * Added a new screen which lets you quickly see all your bookmarks for a given user. * Built user bookmark page in user settings. Moved user settings to it's own lazy loaded module. Removed unneded debouncing from downloader and just used throttleTime instead. * Removed a not-yet implemented tab from series modal * Fixed a bug in clear bookmarks and adjusted icons within anchors to have proper styling |
||
|
Joseph Milazzo
|
b0df67cdda
|
PDF Support + MORE!!!! (#416)
# Added - Added support for PDFs within Kavita. PDFs will open in the Manga reader and you can read through them as images. PDFs are heavier than archives, so they may take longer to open for reading. (Fixes #187) # Changed - Changed: Major change in how Kavita libraries work. Kavita libraries will now allow for mixed media types, that means you can have raw images, archives, epubs, and pdfs all within your Manga library. In the case that the same Series exists between 2 different types of medias, they will be separated and an icon will show to help you identify the types. The correct reader will open regardless of what library you are on. Note: Nightly users need to delete their Raw Images libraries before updating. # Fixed - Fixed: Fixed an issue where checking if a file was modified since last scan always returned true, meaning we would do more I/O than was needed (Fixes #415) - Fixed: There wasn't enough spacing on the top menu bar on the Manga reader - Fixed: Fixed a bug where user preferences dark mode control always showed true, even if you were not using dark mode # Dev stuff - For image extraction, if there is only 1 image we will extract just that, else we will extract only images - Refactored all the Parser code out of the ScannerService into a self contained class. The class should be created for any scans, allowing multiple tasks to run without any chance of cross over. * Fixed indentation for cs files * Fixed an issue where the logic for if a file had been modified or not was not working and always saying modified, meaning we were doing more file I/O than needed. * Implemented the ability to have PDF books. No reader functionality. * Implemented a basic form of scanning for PDF files. Reworked Image based libraries to remove the need to separate in a special library and instead just work within the Manga/Comic library. * Removed the old library types. * Removed some extra code around old raw library types * Fully implemented PDF support into Kavita by using docnet. Removed old libraries we tried that did not work. PDFs take about 200ms to save the file to disk, so they are much slower than reading archives. * Refactored Libraries so that they can have any file extension and the UI will decide which reader to use. * Reworked the Series Parsing code. We now use a separate instance for each task call, so there should be no cross over if 2 tasks are running at the same time. Second, we now store Format with the Series, so we can have duplicate Series with the same name, but a different type of files underneath. * Fixed PDF transparency issues - Used this code to fix an issue when a PDF page doesn't have a background. https://github.com/GowenGit/docnet/issues/8#issuecomment-538985672 - This also fixes the same issue for cover images * Fixed an issue where if a raw image was in a directory with non-image files, those would get moved to cache when trying to open the file. * For image extraction, if there is only 1 image, just copy that to cache instead of everything else in the directory that is an image. * Add some spacing to the top menu bar * Added an icon to the card to showcase the type of file * Added a tag badge to the series detail page * Fixed a bug in user preferences where dark mode control would default to true, even if you weren't on it * Fixed some tests up * Some code smells Co-authored-by: Robbie Davis <robbie@therobbiedavis.com> |
||
|
Joseph Milazzo
|
5083608f24
|
Bugfix/sentry and fixes (#243)
* Generate SeriesMetadata when creating Series from Scanner. * Ignore errors from BookService * Fixed a case where we used First() when it should have been FirstOrDefault() to fail when there are no cover images (or images) |
||
|
Joseph Milazzo
|
d7d7f9b529
|
Collection Support (#234)
* Readme refactored to be more clean and clear, taking inspiration from wiki.js's readme. * Initial backend for Collections and basic metadata implemented. * More build flavors for Raspberry Pi users and updated Install since we don't need users to set their own JWT Token Key. Update a typo in appsettings.json file for prod. * Fixed #224. Sort before getting a First?Last() chatper * The rough ability to add and get series metadata and tags. * Fix a bug on getting metadata for when it doesn't exist. * Fixed a bug where flattening directories with some unique filenames could cause reading order of images to be out of order. * Added a seed code to ensure all series have SeriesMetdata * Ensure all instances of opening an epub is using "using" so we don't lock the file. When we have a malformed html file, log the issues and inform the user we can't open the file. * Book reader now handles @Import "" statements in CSS and inlines the css into css file that references them. This allows for them to be scoped. In addition, if the html or body tag had classes, we now send back a single div with those classes. * Fixed GetSeriesDtoForCollectionAsync which was not properly returning series * Implemented cover image for collection tag. Fixed an issue in metadata update call. * Add check for user access when resolving series for a collection tag. When asking for all tags, if the user is not an admin, only give promotoed tags back. * Implemented updateTag api * Implemented the ability to update series the tags have access to. * Cleanup, sorting, and null check * More sorting changes * Ensure we can delete tags when editing a series tags * Fix order of update to make sure a tag is properly deleted * Code smells |
||
|
Joseph Milazzo
|
a01613f80f
|
EPUB Support (#178)
* Added book filetype detection and reorganized tests due to size of file * Added ability to get basic Parse Info from Book and Pages. * We can now scan books and get them in a library with cover images. * Take the first image in the epub if the cover isn't set. * Implemented the ability to unzip the ebup to cache. Implemented a test api to load html files. * Just some test code to figure out how to approach this. * Fixed some merge conflicts * Removed some dead code from merge * Snapshot: I can now load everything properly into the UI by rewriting the urls before I send them back. I don't notice any lag from this method. It can be optimized further. * Implemented a way to load the content in the browser not via an iframe. * Added a note * Anchor mappings is complete. New anchors are updated so references now resolve to javascript:void() for UI to take care of internally loading and the appropriate page is mapped to it. Anchors that are external have target="_blank" added so they don't force you out of the app and styles are of course inlined. * Oops i need this * Table of contents api implemented (rough) and some small enhancements to codebase for books. * GetBookPageResources now only loads files from within the book. Nested chapter list support and images now use html parsing instead of string parsing. * Fonts now are remapped to load from endpoint. * book-resources now uses a key, ensuring the file is in proper format for lookup. Changed chapter list based on structure with one HEADER and nested chapters. * Properly handle svg resource requests and when there are part anchors that are clickable, make sure we handle them in the UI by adding a kavita-page handler. * Add Chapter group page even if one isn't set by using first page (without part) from nestedChildren. * Added extra debug code for issue #163. * Added new user preferences for books and updated the css so we scope it to our reading section. * Cleaned up style code * Implemented ability to save book preferences and some cleanup on existing apis. * Added an api for checking if a user has read something in a library type before. * Forgot to make sure the has reading progress is against a user lol. * Remove cacheservice code for books, sine we use an in-memory method * Handle svg images as well * Enhanced cover image extraction to check for a "cover" image if the cover image wasn't set in OPF before falling back to the first image. * Fixed an issue with special books not properly generating metadata due to not having filename set. * Cleanup, removed warmup task code from statup/program and changed taskscheduler to schedule tasks on startup only (or if tasks are changed from UI). * Code cleanup * Code cleanup * So much code. Lots of refactors to try to test scanner service. Moved a lot of the queries into Extensions to allow to easier test, even though it's hacky. Support @font-face src:url swaps with ' and ". Source summary information from epubs. * Well...baseURL needs to come from BE and not from UI lol. * Adjusted migrations so default values match Entity * Removed comment * I think I finally fixed #163! The issue was that when i checked if it had a parserInfo, i wasn't considering that the chapter range might have a - in it (0-6) and so when the code to check if range could parse out a number failed, it treated it like a special and checked range against info's filename. * Some bugfixes * Lots of testing, extracting code to make it easier to test. This code is buggy, but fixed a bug where 1) If we changed the normalization code, we would remove the whole db during a scan and 2) We weren't actually removing series properly. Other than that, code is being extracted to remove duplication and centralize logic. * More code cleanup and test cleanup to ensure scan loop is working as expected and matches expectaions from tests. * Cleaned up the code and made it so if I change normalization, which I do in this branch, it wont break existing DBs. * Some comic parser changes for partial chapter support. * Added some code for directory service and scanner service along with python code to generate test files (not used yet). Fixed up all the tests. * Code smells |