Kavita/API/Services/Tasks/Metadata/WordCountAnalyzerService.cs
Joseph Milazzo bbc48a5f5b
Infinite Scroll + List View + Cover Upload Redesign (#1319)
* Started with the redesign of the cover image chooser redesign to be less click intensive for volume/chapter images.

Made some headings bold in card detail drawer.

* Tweaked the styles

* Moved where the info cards show

* Added an ability to open a page settings drawer

* Cleaned up some old code that isn't needed anymore.

* Started implementing a list view. Refactored some title code to a dedicated component

* List view implemented but way too many API calls. Either need caching or adjusting the SeriesDetail api.

* Fixed a bug where if the progress bar didn't render on a card item while a download was in progress, the download indicator would be removed.

* Large refactor to move a lot of the needed fields to the chapter and volume dtos for series detail. All fields are noted when only used in series detail.

* Implemented cards for other tabs (except related)

* Fixed the unit test which needed a mocked reader service call.

* More cleanup around age rating and removing old code from the refactor. Commented out sorting till i feel motivated to work on that.

* Some cleanup and restored cards as initial layout. Time to test this out and see if there is value add.

* Added ability for Chapters tab to show the volume chapters belong to (if applicable)

* Adding style fixes

* Cover image updates, don't allow the first image (which is what is currently set) to respond to cover changes.

Hide the ID field on list item for series detail.

* Refactored the title for list item to be injectable

* Cleaned up the selection code to make it less finicky on mobile when tap scrolling.

* Refactored chapter tab to show volume as well on list view.

* Ensure word count shows for Volumes

* Started adding virtual scrolling, pushing up so Robbie can mess around

* Started adding virtual scrolling, pushing up so Robbie can mess around

* Fixed a bug where all chapters would come under specials

* Show title data as accent if set.

* Style fixes for virtual scroller

* Restyling scroll

* Implemented a way to show storyline with virtual scrolling

* Show Word Count for chapters and cleaned up some logics.

* I might have card layout working with virtual scroll code.

* Some cleanup to hide more system like properties from info bar on series detail page. Fixed some missing time estimate info on storyline chapters.

* Fixed a regression on series service when I integrated VolumeTitle.

* Refactored read time to the backend. Added WordCount to the volume itself so we don't need to calculate on frontend. When asking to analyze files from a series, force the calculation.

* Fixed SeriesDetail api code

* Fixed up the code in the drawer to better update list/card mode

* Basic infinite scroll implemented, however due to how we are updating the list to render, we are re-rending cards that haven't been touched.

* Updated how we render and layout data for infinite scroll on library detail. It's almost there.

* Started laying foundation for loading pages backwards.

Removed lazy loading of images since we are now using virtual paging.

* Hooked in some basic code to allow user to load a prev page with infinite scroll.

* Fixed up series detail api and undid the non-lazy loaded images.

Changed the router to help with this infinite loading on Firefox issue.

* Fixed up some naming issues with Series Detail and added a new test.

* This is an infinite scroll without pagination implementation. It is not fully done, but off to a good start. Virtual scroller with jump bar is working pretty well, def needs more polishing and tweaking. There are hacks in this implementation that need to be revisited.

* Refactored code so that we don't use any pagination and load all results by default.

* Misc code cleanup from build warnings.

* Cleaned up some logic for how to display titles in list view.

* More title cleanup for specials

* Hooked up page layout to user preferences and renamed an existing user pref name to match the dto.

* Swapped out everything but storyline with virtual-scroller over CDK

* Removed CDK from series detail.

* Default value for migration on page layout

* Updating card layout for library detail page

* fixing height for mobile

* Moved scrollbar

* Tweaked some styling for layouts when there is no data

* Refactored the series cards into their own component to make it re-usable.

* More tweaks on series info cards layout and enhanced a few pages with trackby functions.

* Removed some dead code

* Added download on series detail to actionables to fit in with new scroll strategy.

* Fixed language not being updated and sent to the backend for series update.

* Fixed a bad migration (if you ran any prior migration in this branch, you need to undo before you use this commit)

* Adding sticky tabs

* fixed mobile gap on sticky tab

* Enhanced the card title for books to show number up front.

* Adjusted the gutters on admin dashboard

* Removed debug code

* Removing duplicate book title

* Cleaned up old references to cdk scroller

* Implemented a basic jump bar scaling algorithm. Not perfect, but works pretty well.

* Code smells

Co-authored-by: Robbie Davis <robbie@therobbiedavis.com>
2022-06-13 14:37:49 -07:00

241 lines
10 KiB
C#

using System;
using System.Diagnostics;
using System.Linq;
using System.Threading.Tasks;
using API.Data;
using API.Data.Repositories;
using API.Entities;
using API.Entities.Enums;
using API.Helpers;
using API.SignalR;
using Hangfire;
using HtmlAgilityPack;
using Microsoft.Extensions.Logging;
using VersOne.Epub;
namespace API.Services.Tasks.Metadata;
public interface IWordCountAnalyzerService
{
[DisableConcurrentExecution(timeoutInSeconds: 60 * 60 * 60)]
[AutomaticRetry(Attempts = 0, OnAttemptsExceeded = AttemptsExceededAction.Delete)]
Task ScanLibrary(int libraryId, bool forceUpdate = false);
Task ScanSeries(int libraryId, int seriesId, bool forceUpdate = true);
}
/// <summary>
/// This service is a metadata task that generates information around time to read
/// </summary>
public class WordCountAnalyzerService : IWordCountAnalyzerService
{
private readonly ILogger<WordCountAnalyzerService> _logger;
private readonly IUnitOfWork _unitOfWork;
private readonly IEventHub _eventHub;
private readonly ICacheHelper _cacheHelper;
private readonly IReaderService _readerService;
public WordCountAnalyzerService(ILogger<WordCountAnalyzerService> logger, IUnitOfWork unitOfWork, IEventHub eventHub,
ICacheHelper cacheHelper, IReaderService readerService)
{
_logger = logger;
_unitOfWork = unitOfWork;
_eventHub = eventHub;
_cacheHelper = cacheHelper;
_readerService = readerService;
}
public async Task ScanLibrary(int libraryId, bool forceUpdate = false)
{
var sw = Stopwatch.StartNew();
var library = await _unitOfWork.LibraryRepository.GetLibraryForIdAsync(libraryId, LibraryIncludes.None);
await _eventHub.SendMessageAsync(MessageFactory.NotificationProgress,
MessageFactory.WordCountAnalyzerProgressEvent(libraryId, 0F, ProgressEventType.Started, string.Empty));
var chunkInfo = await _unitOfWork.SeriesRepository.GetChunkInfo(library.Id);
var stopwatch = Stopwatch.StartNew();
var totalTime = 0L;
_logger.LogInformation("[MetadataService] Refreshing Library {LibraryName}. Total Items: {TotalSize}. Total Chunks: {TotalChunks} with {ChunkSize} size", library.Name, chunkInfo.TotalSize, chunkInfo.TotalChunks, chunkInfo.ChunkSize);
await _eventHub.SendMessageAsync(MessageFactory.NotificationProgress,
MessageFactory.WordCountAnalyzerProgressEvent(library.Id, 0F, ProgressEventType.Started, $"Starting {library.Name}"));
for (var chunk = 1; chunk <= chunkInfo.TotalChunks; chunk++)
{
if (chunkInfo.TotalChunks == 0) continue;
totalTime += stopwatch.ElapsedMilliseconds;
stopwatch.Restart();
_logger.LogInformation("[MetadataService] Processing chunk {ChunkNumber} / {TotalChunks} with size {ChunkSize}. Series ({SeriesStart} - {SeriesEnd}",
chunk, chunkInfo.TotalChunks, chunkInfo.ChunkSize, chunk * chunkInfo.ChunkSize, (chunk + 1) * chunkInfo.ChunkSize);
var nonLibrarySeries = await _unitOfWork.SeriesRepository.GetFullSeriesForLibraryIdAsync(library.Id,
new UserParams()
{
PageNumber = chunk,
PageSize = chunkInfo.ChunkSize
});
_logger.LogDebug("[MetadataService] Fetched {SeriesCount} series for refresh", nonLibrarySeries.Count);
var seriesIndex = 0;
foreach (var series in nonLibrarySeries)
{
var index = chunk * seriesIndex;
var progress = Math.Max(0F, Math.Min(1F, index * 1F / chunkInfo.TotalSize));
await _eventHub.SendMessageAsync(MessageFactory.NotificationProgress,
MessageFactory.WordCountAnalyzerProgressEvent(library.Id, progress, ProgressEventType.Updated, series.Name));
try
{
await ProcessSeries(series, forceUpdate, false);
}
catch (Exception ex)
{
_logger.LogError(ex, "[MetadataService] There was an exception during metadata refresh for {SeriesName}", series.Name);
}
seriesIndex++;
}
if (_unitOfWork.HasChanges())
{
await _unitOfWork.CommitAsync();
}
_logger.LogInformation(
"[MetadataService] Processed {SeriesStart} - {SeriesEnd} out of {TotalSeries} series in {ElapsedScanTime} milliseconds for {LibraryName}",
chunk * chunkInfo.ChunkSize, (chunk * chunkInfo.ChunkSize) + nonLibrarySeries.Count, chunkInfo.TotalSize, stopwatch.ElapsedMilliseconds, library.Name);
}
await _eventHub.SendMessageAsync(MessageFactory.NotificationProgress,
MessageFactory.WordCountAnalyzerProgressEvent(library.Id, 1F, ProgressEventType.Ended, $"Complete"));
_logger.LogInformation("[WordCountAnalyzerService] Updated metadata for {LibraryName} in {ElapsedMilliseconds} milliseconds", library.Name, sw.ElapsedMilliseconds);
}
public async Task ScanSeries(int libraryId, int seriesId, bool forceUpdate = true)
{
var sw = Stopwatch.StartNew();
var series = await _unitOfWork.SeriesRepository.GetFullSeriesForSeriesIdAsync(seriesId);
if (series == null)
{
_logger.LogError("[WordCountAnalyzerService] Series {SeriesId} was not found on Library {LibraryId}", seriesId, libraryId);
return;
}
await _eventHub.SendMessageAsync(MessageFactory.NotificationProgress,
MessageFactory.WordCountAnalyzerProgressEvent(libraryId, 0F, ProgressEventType.Started, series.Name));
await ProcessSeries(series, forceUpdate);
if (_unitOfWork.HasChanges())
{
await _unitOfWork.CommitAsync();
}
await _eventHub.SendMessageAsync(MessageFactory.NotificationProgress,
MessageFactory.WordCountAnalyzerProgressEvent(libraryId, 1F, ProgressEventType.Ended, series.Name));
_logger.LogInformation("[WordCountAnalyzerService] Updated metadata for {SeriesName} in {ElapsedMilliseconds} milliseconds", series.Name, sw.ElapsedMilliseconds);
}
private async Task ProcessSeries(Series series, bool forceUpdate = false, bool useFileName = true)
{
var isEpub = series.Format == MangaFormat.Epub;
foreach (var volume in series.Volumes)
{
foreach (var chapter in volume.Chapters)
{
// This compares if it's changed since a file scan only
if (!_cacheHelper.HasFileNotChangedSinceCreationOrLastScan(chapter, forceUpdate,
chapter.Files.FirstOrDefault()) && chapter.WordCount != 0)
continue;
if (series.Format == MangaFormat.Epub)
{
long sum = 0;
var fileCounter = 1;
foreach (var file in chapter.Files.Select(file => file.FilePath))
{
var pageCounter = 1;
try
{
using var book = await EpubReader.OpenBookAsync(file, BookService.BookReaderOptions);
var totalPages = book.Content.Html.Values;
foreach (var bookPage in totalPages)
{
var progress = Math.Max(0F,
Math.Min(1F, (fileCounter * pageCounter) * 1F / (chapter.Files.Count * totalPages.Count)));
await _eventHub.SendMessageAsync(MessageFactory.NotificationProgress,
MessageFactory.WordCountAnalyzerProgressEvent(series.LibraryId, progress,
ProgressEventType.Updated, useFileName ? file : series.Name));
sum += await GetWordCountFromHtml(bookPage);
pageCounter++;
}
fileCounter++;
}
catch (Exception ex)
{
_logger.LogError(ex, "There was an error reading an epub file for word count, series skipped");
await _eventHub.SendMessageAsync(MessageFactory.Error,
MessageFactory.ErrorEvent("There was an issue counting words on an epub",
$"{series.Name} - {file}"));
return;
}
}
chapter.WordCount = sum;
series.WordCount += sum;
volume.WordCount += sum;
}
var est = _readerService.GetTimeEstimate(chapter.WordCount, chapter.Pages, isEpub);
chapter.MinHoursToRead = est.MinHours;
chapter.MaxHoursToRead = est.MaxHours;
chapter.AvgHoursToRead = est.AvgHours;
_unitOfWork.ChapterRepository.Update(chapter);
}
var volumeEst = _readerService.GetTimeEstimate(volume.WordCount, volume.Pages, isEpub);
volume.MinHoursToRead = volumeEst.MinHours;
volume.MaxHoursToRead = volumeEst.MaxHours;
volume.AvgHoursToRead = volumeEst.AvgHours;
_unitOfWork.VolumeRepository.Update(volume);
}
var seriesEstimate = _readerService.GetTimeEstimate(series.WordCount, series.Pages, isEpub);
series.MinHoursToRead = seriesEstimate.MinHours;
series.MaxHoursToRead = seriesEstimate.MaxHours;
series.AvgHoursToRead = seriesEstimate.AvgHours;
_unitOfWork.SeriesRepository.Update(series);
}
private static async Task<int> GetWordCountFromHtml(EpubContentFileRef bookFile)
{
var doc = new HtmlDocument();
doc.LoadHtml(await bookFile.ReadContentAsTextAsync());
var textNodes = doc.DocumentNode.SelectNodes("//body//text()[not(parent::script)]");
if (textNodes == null) return 0;
return textNodes
.Select(node => node.InnerText.Split(' ', StringSplitOptions.RemoveEmptyEntries)
.Where(s => char.IsLetter(s[0])))
.Select(words => words.Count())
.Where(wordCount => wordCount > 0)
.Sum();
}
}