Kavita/API/Services/Tasks/Scanner/ParseScannedFiles.cs
Joseph Milazzo 4e49aa47ce
Change Detection: On Push aka UI Smoothness (#1369)
* Updated Series Info Cards to use OnPush and hooked in progress events when we do a mark as read/unread on entities. These events update progress bars but also will now trigger a re-calculation on Read Time Left.

* Removed Library Card Component

* Refactored manga reader title and subtitle calculation to the backend.

* Coverted card actionables to onPush

* Series Card on push cleanup

* Updated edit collection tags for on push

* Update cover image chooser for on push

* Cleaned up carsouel reel

* Updated cover image to allow for uploading gif and webp files

* Bulk add to collection on push

* Updated bulk operation to use on push. Updated bulk operation to have mark as unread and read buttons explicitly. Updated so add to collection is visible and delete.

Fixed a bug where manage library component wasn't invoking the trackBy function

* Updating entity title for on push

* Removed file info component

* Updated Mange Library for on push

* Entity info cards on push

* List item on push

* Updated icon and title for on push and fixed some missing change detection on series detail

* Restricted the typeahead interface to simplify the design

* Edit Series Relation now shows a value in the dropdown for Parent relationships and disables the field.

* Updated edit series relation to focus on new typeahead when adding a new relationship

* Added some documentation and when Scanning a library, don't allow the user to enqueue the same job multiple times.

* Applied the No-enqueue if already enqueued logic to other tasks

* Library detail on push

* Updated events widget to onpush

* Card detail drawer on push. Card detail cover chooser now will show all chapter's covers for selection in cover chooser.

* Chapter metadata detail on push

* Removed Card Detail modal

* All collections on push

* Removed some comments

* Updated bulk selection to use an observable rather than function calls so new on push strategy works

* collection detail now uses on push and scroller is placed on correct element

* Updated library recommended to on push. Ensure that when mark as read occurs, the appropriate streams are refreshed.

* Updated library detail to on push

* Update metadata fiter to onpush. Bugs found and reported to Project

* person badge on push

* Read more on push

* Updated tag badge to on push

* User login on push

* When initing side nav, don't call an authenticated api until we are sure a user is logged in

* Updated splash container to on push

* Dashboard on push

* Side nav slight refactor around some api calls

* Cleaned up series card on push to use same cdRef naming convention

* Updated Static Files to use caching

* Added width and height to logo image

* shortcuts modal on push

* reading lists on push

* Reading list detail on push

* draggable ordered list on push

* Refactored reading-list-detail to use a new item which drastically reduces renders on operations

* series format on push

* circular loader on push

* Badge Expander on push

* update notification modal on push

* drawer on push

* Edit Series Modal on push

* reset password on push

* review series modal on push

* series metadata detail on push

* theme manager on push

* confirm reset password on push

* register on push

* confirm migration email on push

* confirm email on push

* add email to account migration on push

* user preferences on push. Made global settings default open

* edit series relation on push

* Fixed an edge case bug for next chapter where if the current volume had a single chapter of 1 and the next volume had a chapter number of 0, it would say there are no more chapters.

* Updated infinite scroller with on push support

* Moved some animations over to typeahead, not integrated yet.

* Manga reader is now on push

* Reader settings on push

* refactored how we close the book

* Updated table of contents for on push

* Updated book reader for on push. Fixed a bug where table of contents wasn't showing current page anchor due to a scroll calulation bug

* Small code tweak

* Icon and title on push

* nav header on push

* grouped typeahead on push

* typeahead on push and added a new trackby identity function to allow even faster rendering of big lists

* pdf reader on push

* code cleanup
2022-07-11 08:57:07 -07:00

302 lines
13 KiB
C#

using System;
using System.Collections.Concurrent;
using System.Collections.Generic;
using System.Diagnostics;
using System.IO;
using System.Linq;
using System.Threading.Tasks;
using API.Data.Metadata;
using API.Entities;
using API.Entities.Enums;
using API.Helpers;
using API.Parser;
using API.SignalR;
using Microsoft.AspNetCore.SignalR;
using Microsoft.Extensions.Logging;
namespace API.Services.Tasks.Scanner
{
public class ParsedSeries
{
public string Name { get; init; }
public string NormalizedName { get; init; }
public MangaFormat Format { get; init; }
}
public class ParseScannedFiles
{
private readonly ConcurrentDictionary<ParsedSeries, List<ParserInfo>> _scannedSeries;
private readonly ILogger _logger;
private readonly IDirectoryService _directoryService;
private readonly IReadingItemService _readingItemService;
private readonly IEventHub _eventHub;
private readonly DefaultParser _defaultParser;
/// <summary>
/// An instance of a pipeline for processing files and returning a Map of Series -> ParserInfos.
/// Each instance is separate from other threads, allowing for no cross over.
/// </summary>
/// <param name="logger">Logger of the parent class that invokes this</param>
/// <param name="directoryService">Directory Service</param>
/// <param name="readingItemService">ReadingItemService Service for extracting information on a number of formats</param>
/// <param name="eventHub">For firing off SignalR events</param>
public ParseScannedFiles(ILogger logger, IDirectoryService directoryService,
IReadingItemService readingItemService, IEventHub eventHub)
{
_logger = logger;
_directoryService = directoryService;
_readingItemService = readingItemService;
_scannedSeries = new ConcurrentDictionary<ParsedSeries, List<ParserInfo>>();
_defaultParser = new DefaultParser(_directoryService);
_eventHub = eventHub;
}
/// <summary>
/// Gets the list of all parserInfos given a Series (Will match on Name, LocalizedName, OriginalName). If the series does not exist within, return empty list.
/// </summary>
/// <param name="parsedSeries"></param>
/// <param name="series"></param>
/// <returns></returns>
public static IList<ParserInfo> GetInfosByName(Dictionary<ParsedSeries, List<ParserInfo>> parsedSeries, Series series)
{
var allKeys = parsedSeries.Keys.Where(ps =>
SeriesHelper.FindSeries(series, ps));
var infos = new List<ParserInfo>();
foreach (var key in allKeys)
{
infos.AddRange(parsedSeries[key]);
}
return infos;
}
/// <summary>
/// Processes files found during a library scan.
/// Populates a collection of <see cref="ParserInfo"/> for DB updates later.
/// </summary>
/// <param name="path">Path of a file</param>
/// <param name="rootPath"></param>
/// <param name="type">Library type to determine parsing to perform</param>
private void ProcessFile(string path, string rootPath, LibraryType type)
{
var info = _readingItemService.Parse(path, rootPath, type);
if (info == null)
{
// If the file is an image and literally a cover image, skip processing.
if (!(Parser.Parser.IsImage(path) && Parser.Parser.IsCoverImage(path)))
{
_logger.LogWarning("[Scanner] Could not parse series from {Path}", path);
}
return;
}
// This catches when original library type is Manga/Comic and when parsing with non
if (Parser.Parser.IsEpub(path) && Parser.Parser.ParseVolume(info.Series) != Parser.Parser.DefaultVolume) // Shouldn't this be info.Volume != DefaultVolume?
{
info = _defaultParser.Parse(path, rootPath, LibraryType.Book);
var info2 = _readingItemService.Parse(path, rootPath, type);
info.Merge(info2);
}
info.ComicInfo = _readingItemService.GetComicInfo(path);
if (info.ComicInfo != null)
{
if (!string.IsNullOrEmpty(info.ComicInfo.Volume))
{
info.Volumes = info.ComicInfo.Volume;
}
if (!string.IsNullOrEmpty(info.ComicInfo.Series))
{
info.Series = info.ComicInfo.Series.Trim();
}
if (!string.IsNullOrEmpty(info.ComicInfo.Number))
{
info.Chapters = info.ComicInfo.Number;
}
// Patch is SeriesSort from ComicInfo
if (!string.IsNullOrEmpty(info.ComicInfo.TitleSort))
{
info.SeriesSort = info.ComicInfo.TitleSort.Trim();
}
if (!string.IsNullOrEmpty(info.ComicInfo.Format) && Parser.Parser.HasComicInfoSpecial(info.ComicInfo.Format))
{
info.IsSpecial = true;
info.Chapters = Parser.Parser.DefaultChapter;
info.Volumes = Parser.Parser.DefaultVolume;
}
if (!string.IsNullOrEmpty(info.ComicInfo.SeriesSort))
{
info.SeriesSort = info.ComicInfo.SeriesSort.Trim();
}
if (!string.IsNullOrEmpty(info.ComicInfo.LocalizedSeries))
{
info.LocalizedSeries = info.ComicInfo.LocalizedSeries.Trim();
}
}
try
{
TrackSeries(info);
}
catch (Exception ex)
{
_logger.LogError(ex, "There was an exception that occurred during tracking {FilePath}. Skipping this file", info.FullFilePath);
}
}
/// <summary>
/// Attempts to either add a new instance of a show mapping to the _scannedSeries bag or adds to an existing.
/// This will check if the name matches an existing series name (multiple fields) <see cref="MergeName"/>
/// </summary>
/// <param name="info"></param>
private void TrackSeries(ParserInfo info)
{
if (info.Series == string.Empty) return;
// Check if normalized info.Series already exists and if so, update info to use that name instead
info.Series = MergeName(info);
var normalizedSeries = Parser.Parser.Normalize(info.Series);
var normalizedSortSeries = Parser.Parser.Normalize(info.SeriesSort);
var normalizedLocalizedSeries = Parser.Parser.Normalize(info.LocalizedSeries);
try
{
var existingKey = _scannedSeries.Keys.SingleOrDefault(ps =>
ps.Format == info.Format && (ps.NormalizedName.Equals(normalizedSeries)
|| ps.NormalizedName.Equals(normalizedLocalizedSeries)
|| ps.NormalizedName.Equals(normalizedSortSeries)));
existingKey ??= new ParsedSeries()
{
Format = info.Format,
Name = info.Series,
NormalizedName = normalizedSeries
};
_scannedSeries.AddOrUpdate(existingKey, new List<ParserInfo>() {info}, (_, oldValue) =>
{
oldValue ??= new List<ParserInfo>();
if (!oldValue.Contains(info))
{
oldValue.Add(info);
}
return oldValue;
});
}
catch (Exception ex)
{
_logger.LogCritical(ex, "{SeriesName} matches against multiple series in the parsed series. This indicates a critical kavita issue. Key will be skipped", info.Series);
foreach (var seriesKey in _scannedSeries.Keys.Where(ps =>
ps.Format == info.Format && (ps.NormalizedName.Equals(normalizedSeries)
|| ps.NormalizedName.Equals(normalizedLocalizedSeries)
|| ps.NormalizedName.Equals(normalizedSortSeries))))
{
_logger.LogCritical("Matches: {SeriesName} matches on {SeriesKey}", info.Series, seriesKey.Name);
}
}
}
/// <summary>
/// Using a normalized name from the passed ParserInfo, this checks against all found series so far and if an existing one exists with
/// same normalized name, it merges into the existing one. This is important as some manga may have a slight difference with punctuation or capitalization.
/// </summary>
/// <param name="info"></param>
/// <returns>Series Name to group this info into</returns>
public string MergeName(ParserInfo info)
{
var normalizedSeries = Parser.Parser.Normalize(info.Series);
var normalizedLocalSeries = Parser.Parser.Normalize(info.LocalizedSeries);
// We use FirstOrDefault because this was introduced late in development and users might have 2 series with both names
try
{
var existingName =
_scannedSeries.SingleOrDefault(p =>
(Parser.Parser.Normalize(p.Key.NormalizedName) == normalizedSeries ||
Parser.Parser.Normalize(p.Key.NormalizedName) == normalizedLocalSeries) &&
p.Key.Format == info.Format)
.Key;
if (existingName != null && !string.IsNullOrEmpty(existingName.Name))
{
return existingName.Name;
}
}
catch (Exception ex)
{
_logger.LogCritical(ex, "Multiple series detected for {SeriesName} ({File})! This is critical to fix! There should only be 1", info.Series, info.FullFilePath);
var values = _scannedSeries.Where(p =>
(Parser.Parser.Normalize(p.Key.NormalizedName) == normalizedSeries ||
Parser.Parser.Normalize(p.Key.NormalizedName) == normalizedLocalSeries) &&
p.Key.Format == info.Format);
foreach (var pair in values)
{
_logger.LogCritical("Duplicate Series in DB matches with {SeriesName}: {DuplicateName}", info.Series, pair.Key.Name);
}
}
return info.Series;
}
/// <summary>
///
/// </summary>
/// <param name="libraryType">Type of library. Used for selecting the correct file extensions to search for and parsing files</param>
/// <param name="folders">The folders to scan. By default, this should be library.Folders, however it can be overwritten to restrict folders</param>
/// <param name="libraryName">Name of the Library</param>
/// <returns></returns>
public async Task<Dictionary<ParsedSeries, List<ParserInfo>>> ScanLibrariesForSeries(LibraryType libraryType, IEnumerable<string> folders, string libraryName)
{
await _eventHub.SendMessageAsync(MessageFactory.NotificationProgress, MessageFactory.FileScanProgressEvent("", libraryName, ProgressEventType.Started));
foreach (var folderPath in folders)
{
try
{
async void Action(string f)
{
try
{
ProcessFile(f, folderPath, libraryType);
await _eventHub.SendMessageAsync(MessageFactory.NotificationProgress, MessageFactory.FileScanProgressEvent(f, libraryName, ProgressEventType.Updated));
}
catch (FileNotFoundException exception)
{
_logger.LogError(exception, "The file {Filename} could not be found", f);
}
}
_directoryService.TraverseTreeParallelForEach(folderPath, Action, Parser.Parser.SupportedExtensions, _logger);
}
catch (ArgumentException ex)
{
_logger.LogError(ex, "The directory '{FolderPath}' does not exist", folderPath);
}
}
await _eventHub.SendMessageAsync(MessageFactory.NotificationProgress, MessageFactory.FileScanProgressEvent("", libraryName, ProgressEventType.Ended));
return SeriesWithInfos();
}
/// <summary>
/// Returns any series where there were parsed infos
/// </summary>
/// <returns></returns>
private Dictionary<ParsedSeries, List<ParserInfo>> SeriesWithInfos()
{
var filtered = _scannedSeries.Where(kvp => kvp.Value.Count > 0);
var series = filtered.ToDictionary(v => v.Key, v => v.Value);
return series;
}
}
}