using System; using System.Collections.Generic; using System.Diagnostics; using System.IO; using System.IO.Compression; using System.Linq; using System.Text; using System.Threading.Tasks; using System.Xml.Serialization; using API.Archive; using API.Data.Metadata; using API.Extensions; using API.Services.Tasks; using Kavita.Common; using Microsoft.Extensions.Logging; using SharpCompress.Archives; using SharpCompress.Common; namespace API.Services { public interface IArchiveService { void ExtractArchive(string archivePath, string extractPath); int GetNumberOfPagesFromArchive(string archivePath); string GetCoverImage(string archivePath, string fileName, string outputDirectory); bool IsValidArchive(string archivePath); ComicInfo GetComicInfo(string archivePath); ArchiveLibrary CanOpen(string archivePath); bool ArchiveNeedsFlattening(ZipArchive archive); Task> CreateZipForDownload(IEnumerable files, string tempFolder); } /// /// Responsible for manipulating Archive files. Used by and /// // ReSharper disable once ClassWithVirtualMembersNeverInherited.Global public class ArchiveService : IArchiveService { private readonly ILogger _logger; private readonly IDirectoryService _directoryService; private readonly IImageService _imageService; private const string ComicInfoFilename = "comicinfo"; public ArchiveService(ILogger logger, IDirectoryService directoryService, IImageService imageService) { _logger = logger; _directoryService = directoryService; _imageService = imageService; } /// /// Checks if a File can be opened. Requires up to 2 opens of the filestream. /// /// /// public virtual ArchiveLibrary CanOpen(string archivePath) { if (string.IsNullOrEmpty(archivePath) || !(File.Exists(archivePath) && Parser.Parser.IsArchive(archivePath) || Parser.Parser.IsEpub(archivePath))) return ArchiveLibrary.NotSupported; var ext = _directoryService.FileSystem.Path.GetExtension(archivePath).ToUpper(); if (ext.Equals(".CBR") || ext.Equals(".RAR")) return ArchiveLibrary.SharpCompress; try { using var a2 = ZipFile.OpenRead(archivePath); return ArchiveLibrary.Default; } catch (Exception) { try { using var a1 = ArchiveFactory.Open(archivePath); return ArchiveLibrary.SharpCompress; } catch (Exception) { return ArchiveLibrary.NotSupported; } } } public int GetNumberOfPagesFromArchive(string archivePath) { if (!IsValidArchive(archivePath)) { _logger.LogError("Archive {ArchivePath} could not be found", archivePath); return 0; } try { var libraryHandler = CanOpen(archivePath); switch (libraryHandler) { case ArchiveLibrary.Default: { using var archive = ZipFile.OpenRead(archivePath); return archive.Entries.Count(e => !Parser.Parser.HasBlacklistedFolderInPath(e.FullName) && Parser.Parser.IsImage(e.FullName)); } case ArchiveLibrary.SharpCompress: { using var archive = ArchiveFactory.Open(archivePath); return archive.Entries.Count(entry => !entry.IsDirectory && !Parser.Parser.HasBlacklistedFolderInPath(Path.GetDirectoryName(entry.Key) ?? string.Empty) && Parser.Parser.IsImage(entry.Key)); } case ArchiveLibrary.NotSupported: _logger.LogWarning("[GetNumberOfPagesFromArchive] This archive cannot be read: {ArchivePath}. Defaulting to 0 pages", archivePath); return 0; default: _logger.LogWarning("[GetNumberOfPagesFromArchive] There was an exception when reading archive stream: {ArchivePath}. Defaulting to 0 pages", archivePath); return 0; } } catch (Exception ex) { _logger.LogWarning(ex, "[GetNumberOfPagesFromArchive] There was an exception when reading archive stream: {ArchivePath}. Defaulting to 0 pages", archivePath); return 0; } } /// /// Finds the first instance of a folder entry and returns it /// /// /// Entry name of match, null if no match public static string FindFolderEntry(IEnumerable entryFullNames) { var result = entryFullNames .Where(path => !(Path.EndsInDirectorySeparator(path) || Parser.Parser.HasBlacklistedFolderInPath(path) || path.StartsWith(Parser.Parser.MacOsMetadataFileStartsWith))) .OrderByNatural(Path.GetFileNameWithoutExtension) .FirstOrDefault(Parser.Parser.IsCoverImage); return string.IsNullOrEmpty(result) ? null : result; } /// /// Returns first entry that is an image and is not in a blacklisted folder path. Uses for ordering files /// /// /// Entry name of match, null if no match public static string? FirstFileEntry(IEnumerable entryFullNames, string archiveName) { // First check if there are any files that are not in a nested folder before just comparing by filename. This is needed // because NaturalSortComparer does not work with paths and doesn't seem 001.jpg as before chapter 1/001.jpg. var fullNames = entryFullNames .Where(path => !(Path.EndsInDirectorySeparator(path) || Parser.Parser.HasBlacklistedFolderInPath(path) || path.StartsWith(Parser.Parser.MacOsMetadataFileStartsWith)) && Parser.Parser.IsImage(path)) .OrderByNatural(c => c.GetFullPathWithoutExtension()) .ToList(); if (fullNames.Count == 0) return null; var nonNestedFile = fullNames.Where(entry => (Path.GetDirectoryName(entry) ?? string.Empty).Equals(archiveName)) .OrderByNatural(c => c.GetFullPathWithoutExtension()) .FirstOrDefault(); if (!string.IsNullOrEmpty(nonNestedFile)) return nonNestedFile; // Check the first folder and sort within that to see if we can find a file, else fallback to first file with basic sort. // Get first folder, then sort within that var firstDirectoryFile = fullNames.OrderByNatural(Path.GetDirectoryName).FirstOrDefault(); if (!string.IsNullOrEmpty(firstDirectoryFile)) { var firstDirectory = Path.GetDirectoryName(firstDirectoryFile); if (!string.IsNullOrEmpty(firstDirectory)) { var firstDirectoryResult = fullNames.Where(f => firstDirectory.Equals(Path.GetDirectoryName(f))) .OrderByNatural(Path.GetFileNameWithoutExtension) .FirstOrDefault(); if (!string.IsNullOrEmpty(firstDirectoryResult)) return firstDirectoryResult; } } var result = fullNames .OrderByNatural(Path.GetFileNameWithoutExtension) .FirstOrDefault(); return string.IsNullOrEmpty(result) ? null : result; } /// /// Generates byte array of cover image. /// Given a path to a compressed file , will ensure the first image (respects directory structure) is returned unless /// a folder/cover.(image extension) exists in the the compressed file (if duplicate, the first is chosen) /// /// This skips over any __MACOSX folder/file iteration. /// /// This always creates a thumbnail /// /// File name to use based on context of entity. /// public string GetCoverImage(string archivePath, string fileName, string outputDirectory) { if (archivePath == null || !IsValidArchive(archivePath)) return string.Empty; try { var libraryHandler = CanOpen(archivePath); switch (libraryHandler) { case ArchiveLibrary.Default: { using var archive = ZipFile.OpenRead(archivePath); var entryName = FindCoverImageFilename(archivePath, archive.Entries.Select(e => e.FullName)); var entry = archive.Entries.Single(e => e.FullName == entryName); using var stream = entry.Open(); return _imageService.WriteCoverThumbnail(stream, fileName, outputDirectory); } case ArchiveLibrary.SharpCompress: { using var archive = ArchiveFactory.Open(archivePath); var entryNames = archive.Entries.Where(archiveEntry => !archiveEntry.IsDirectory).Select(e => e.Key).ToList(); var entryName = FindCoverImageFilename(archivePath, entryNames); var entry = archive.Entries.Single(e => e.Key == entryName); using var stream = entry.OpenEntryStream(); return _imageService.WriteCoverThumbnail(stream, fileName, outputDirectory); } case ArchiveLibrary.NotSupported: _logger.LogWarning("[GetCoverImage] This archive cannot be read: {ArchivePath}. Defaulting to no cover image", archivePath); return string.Empty; default: _logger.LogWarning("[GetCoverImage] There was an exception when reading archive stream: {ArchivePath}. Defaulting to no cover image", archivePath); return string.Empty; } } catch (Exception ex) { _logger.LogWarning(ex, "[GetCoverImage] There was an exception when reading archive stream: {ArchivePath}. Defaulting to no cover image", archivePath); } return string.Empty; } /// /// Given a list of image paths (assume within an archive), find the filename that corresponds to the cover /// /// /// /// public static string FindCoverImageFilename(string archivePath, IEnumerable entryNames) { var entryName = FindFolderEntry(entryNames) ?? FirstFileEntry(entryNames, Path.GetFileName(archivePath)); return entryName; } /// /// Given an archive stream, will assess whether directory needs to be flattened so that the extracted archive files are directly /// under extract path and not nested in subfolders. See Flatten method. /// /// An opened archive stream /// public bool ArchiveNeedsFlattening(ZipArchive archive) { // Sometimes ZipArchive will list the directory and others it will just keep it in the FullName return archive.Entries.Count > 0 && !Path.HasExtension(archive.Entries.ElementAt(0).FullName) || archive.Entries.Any(e => e.FullName.Contains(Path.AltDirectorySeparatorChar) && !Parser.Parser.HasBlacklistedFolderInPath(e.FullName)); } // TODO: Refactor CreateZipForDownload to return the temp file so we can stream it from temp /// /// /// /// /// Temp folder name to use for preparing the files. Will be created and deleted /// /// public async Task> CreateZipForDownload(IEnumerable files, string tempFolder) { var dateString = DateTime.Now.ToShortDateString().Replace("/", "_"); var tempLocation = Path.Join(_directoryService.TempDirectory, $"{tempFolder}_{dateString}"); _directoryService.ExistOrCreate(tempLocation); if (!_directoryService.CopyFilesToDirectory(files, tempLocation)) { throw new KavitaException("Unable to copy files to temp directory archive download."); } var zipPath = Path.Join(_directoryService.TempDirectory, $"kavita_{tempFolder}_{dateString}.zip"); try { ZipFile.CreateFromDirectory(tempLocation, zipPath); } catch (AggregateException ex) { _logger.LogError(ex, "There was an issue creating temp archive"); throw new KavitaException("There was an issue creating temp archive"); } var fileBytes = await _directoryService.ReadFileAsync(zipPath); _directoryService.ClearAndDeleteDirectory(tempLocation); // NOTE: For sending back just zip, just schedule this to be called after the file is returned or let next temp storage cleanup take care of it (new FileInfo(zipPath)).Delete(); return Tuple.Create(fileBytes, zipPath); } /// /// Test if the archive path exists and an archive /// /// /// public bool IsValidArchive(string archivePath) { if (!File.Exists(archivePath)) { _logger.LogWarning("Archive {ArchivePath} could not be found", archivePath); return false; } if (Parser.Parser.IsArchive(archivePath) || Parser.Parser.IsEpub(archivePath)) return true; _logger.LogWarning("Archive {ArchivePath} is not a valid archive", archivePath); return false; } private static bool ValidComicInfoArchiveEntry(string fullName, string name) { var filenameWithoutExtension = Path.GetFileNameWithoutExtension(name).ToLower(); return !Parser.Parser.HasBlacklistedFolderInPath(fullName) && filenameWithoutExtension.Equals(ComicInfoFilename, StringComparison.InvariantCultureIgnoreCase) && !filenameWithoutExtension.StartsWith(Parser.Parser.MacOsMetadataFileStartsWith) && Parser.Parser.IsXml(name); } /// /// This can be null if nothing is found or any errors occur during access /// /// /// public ComicInfo? GetComicInfo(string archivePath) { if (!IsValidArchive(archivePath)) return null; try { if (!File.Exists(archivePath)) return null; var libraryHandler = CanOpen(archivePath); switch (libraryHandler) { case ArchiveLibrary.Default: { using var archive = ZipFile.OpenRead(archivePath); var entry = archive.Entries.FirstOrDefault(x => ValidComicInfoArchiveEntry(x.FullName, x.Name)); if (entry != null) { using var stream = entry.Open(); var serializer = new XmlSerializer(typeof(ComicInfo)); var info = (ComicInfo) serializer.Deserialize(stream); ComicInfo.CleanComicInfo(info); return info; } break; } case ArchiveLibrary.SharpCompress: { using var archive = ArchiveFactory.Open(archivePath); var entry = archive.Entries.FirstOrDefault(entry => ValidComicInfoArchiveEntry(Path.GetDirectoryName(entry.Key), entry.Key)); if (entry != null) { using var stream = entry.OpenEntryStream(); var serializer = new XmlSerializer(typeof(ComicInfo)); var info = (ComicInfo) serializer.Deserialize(stream); ComicInfo.CleanComicInfo(info); return info; } break; } case ArchiveLibrary.NotSupported: _logger.LogWarning("[GetComicInfo] This archive cannot be read: {ArchivePath}", archivePath); return null; default: _logger.LogWarning( "[GetComicInfo] There was an exception when reading archive stream: {ArchivePath}", archivePath); return null; } } catch (Exception ex) { _logger.LogWarning(ex, "[GetComicInfo] There was an exception when reading archive stream: {Filepath}", archivePath); } return null; } private void ExtractArchiveEntities(IEnumerable entries, string extractPath) { _directoryService.ExistOrCreate(extractPath); foreach (var entry in entries) { entry.WriteToDirectory(extractPath, new ExtractionOptions() { ExtractFullPath = true, // Don't flatten, let the flatterner ensure correct order of nested folders Overwrite = false }); } } private void ExtractArchiveEntries(ZipArchive archive, string extractPath) { // TODO: In cases where we try to extract, but there are InvalidPathChars, we need to inform the user (throw exception, let middleware inform user) var needsFlattening = ArchiveNeedsFlattening(archive); if (!archive.HasFiles() && !needsFlattening) return; archive.ExtractToDirectory(extractPath, true); if (!needsFlattening) return; _logger.LogDebug("Extracted archive is nested in root folder, flattening..."); _directoryService.Flatten(extractPath); } /// /// Extracts an archive to a temp cache directory. Returns path to new directory. If temp cache directory already exists, /// will return that without performing an extraction. Returns empty string if there are any invalidations which would /// prevent operations to perform correctly (missing archivePath file, empty archive, etc). /// /// A valid file to an archive file. /// Path to extract to /// public void ExtractArchive(string archivePath, string extractPath) { if (!IsValidArchive(archivePath)) return; if (Directory.Exists(extractPath)) return; if (!_directoryService.FileSystem.File.Exists(archivePath)) { _logger.LogError("{Archive} does not exist on disk", archivePath); throw new KavitaException($"{archivePath} does not exist on disk"); } var sw = Stopwatch.StartNew(); try { var libraryHandler = CanOpen(archivePath); switch (libraryHandler) { case ArchiveLibrary.Default: { using var archive = ZipFile.OpenRead(archivePath); ExtractArchiveEntries(archive, extractPath); break; } case ArchiveLibrary.SharpCompress: { using var archive = ArchiveFactory.Open(archivePath); ExtractArchiveEntities(archive.Entries.Where(entry => !entry.IsDirectory && !Parser.Parser.HasBlacklistedFolderInPath(Path.GetDirectoryName(entry.Key) ?? string.Empty) && Parser.Parser.IsImage(entry.Key)), extractPath); break; } case ArchiveLibrary.NotSupported: _logger.LogWarning("[ExtractArchive] This archive cannot be read: {ArchivePath}", archivePath); return; default: _logger.LogWarning("[ExtractArchive] There was an exception when reading archive stream: {ArchivePath}", archivePath); return; } } catch (Exception e) { _logger.LogWarning(e, "[ExtractArchive] There was a problem extracting {ArchivePath} to {ExtractPath}",archivePath, extractPath); return; } _logger.LogDebug("Extracted archive to {ExtractPath} in {ElapsedMilliseconds} milliseconds", extractPath, sw.ElapsedMilliseconds); } } }