Cleanup extracted files (#13760)

* Cleanup extracted files

* Pagination and fixes

* Add migration for attachments to MigrateLibraryDb

* Unify attachment handling

* Don't extract again if files were already extracted

* Fix MKS attachment extraction

* Always run full extraction on mks

* Don't try to extract mjpeg streams as attachments

* Fallback to check if attachments were extracted to cache folder

* Fixup

3 files changed, 65 insertions, 22 deletions

diff --git a/MediaBrowser.Controller/IO/IPathManager.cs b/MediaBrowser.Controller/IO/IPathManager.cs
index 036889810..7c20164a6 100644
--- a/MediaBrowser.Controller/IO/IPathManager.cs
+++ b/MediaBrowser.Controller/IO/IPathManager.cs
@@ -1,4 +1,5 @@
 using MediaBrowser.Controller.Entities;
+using MediaBrowser.Model.Dto;
 
 namespace MediaBrowser.Controller.IO;
 
@@ -14,4 +15,35 @@ public interface IPathManager
     /// <param name="saveWithMedia">Whether or not the tile should be saved next to the media file.</param>
     /// <returns>The absolute path.</returns>
     public string GetTrickplayDirectory(BaseItem item, bool saveWithMedia = false);
+
+    /// <summary>
+    /// Gets the path to the subtitle file.
+    /// </summary>
+    /// <param name="mediaSourceId">The media source id.</param>
+    /// <param name="streamIndex">The stream index.</param>
+    /// <param name="extension">The subtitle file extension.</param>
+    /// <returns>The absolute path.</returns>
+    public string GetSubtitlePath(string mediaSourceId, int streamIndex, string extension);
+
+    /// <summary>
+    /// Gets the path to the subtitle file.
+    /// </summary>
+    /// <param name="mediaSourceId">The media source id.</param>
+    /// <returns>The absolute path.</returns>
+    public string GetSubtitleFolderPath(string mediaSourceId);
+
+    /// <summary>
+    /// Gets the path to the attachment file.
+    /// </summary>
+    /// <param name="mediaSourceId">The media source id.</param>
+    /// <param name="fileName">The attachmentFileName index.</param>
+    /// <returns>The absolute path.</returns>
+    public string GetAttachmentPath(string mediaSourceId, string fileName);
+
+    /// <summary>
+    /// Gets the path to the attachment folder.
+    /// </summary>
+    /// <param name="mediaSourceId">The media source id.</param>
+    /// <returns>The absolute path.</returns>
+    public string GetAttachmentFolderPath(string mediaSourceId);
 }
diff --git a/MediaBrowser.Controller/MediaEncoding/EncodingHelper.cs b/MediaBrowser.Controller/MediaEncoding/EncodingHelper.cs
index afa962a41..75b3f151d 100644
--- a/MediaBrowser.Controller/MediaEncoding/EncodingHelper.cs
+++ b/MediaBrowser.Controller/MediaEncoding/EncodingHelper.cs
@@ -19,6 +19,7 @@ using Jellyfin.Database.Implementations.Enums;
 using Jellyfin.Extensions;
 using MediaBrowser.Common.Configuration;
 using MediaBrowser.Controller.Extensions;
+using MediaBrowser.Controller.IO;
 using MediaBrowser.Model.Configuration;
 using MediaBrowser.Model.Dlna;
 using MediaBrowser.Model.Dto;
@@ -55,6 +56,7 @@ namespace MediaBrowser.Controller.MediaEncoding
         private readonly ISubtitleEncoder _subtitleEncoder;
         private readonly IConfiguration _config;
         private readonly IConfigurationManager _configurationManager;
+        private readonly IPathManager _pathManager;
 
         // i915 hang was fixed by linux 6.2 (3f882f2)
         private readonly Version _minKerneli915Hang = new Version(5, 18);
@@ -153,13 +155,15 @@ namespace MediaBrowser.Controller.MediaEncoding
             IMediaEncoder mediaEncoder,
             ISubtitleEncoder subtitleEncoder,
             IConfiguration config,
-            IConfigurationManager configurationManager)
+            IConfigurationManager configurationManager,
+            IPathManager pathManager)
         {
             _appPaths = appPaths;
             _mediaEncoder = mediaEncoder;
             _subtitleEncoder = subtitleEncoder;
             _config = config;
             _configurationManager = configurationManager;
+            _pathManager = pathManager;
         }
 
         private enum DynamicHdrMetadataRemovalPlan
@@ -1785,7 +1789,7 @@ namespace MediaBrowser.Controller.MediaEncoding
             var alphaParam = enableAlpha ? ":alpha=1" : string.Empty;
             var sub2videoParam = enableSub2video ? ":sub2video=1" : string.Empty;
 
-            var fontPath = Path.Combine(_appPaths.CachePath, "attachments", state.MediaSource.Id);
+            var fontPath = _pathManager.GetAttachmentFolderPath(state.MediaSource.Id);
             var fontParam = string.Format(
                 CultureInfo.InvariantCulture,
                 ":fontsdir='{0}'",
diff --git a/MediaBrowser.Controller/MediaEncoding/IAttachmentExtractor.cs b/MediaBrowser.Controller/MediaEncoding/IAttachmentExtractor.cs
index 09840d2ee..d8d136472 100644
--- a/MediaBrowser.Controller/MediaEncoding/IAttachmentExtractor.cs
+++ b/MediaBrowser.Controller/MediaEncoding/IAttachmentExtractor.cs
@@ -9,26 +9,33 @@ using MediaBrowser.Controller.Entities;
 using MediaBrowser.Model.Dto;
 using MediaBrowser.Model.Entities;
 
-namespace MediaBrowser.Controller.MediaEncoding
-{
-    public interface IAttachmentExtractor
-    {
-        Task<(MediaAttachment Attachment, Stream Stream)> GetAttachment(
-            BaseItem item,
-            string mediaSourceId,
-            int attachmentStreamIndex,
-            CancellationToken cancellationToken);
+namespace MediaBrowser.Controller.MediaEncoding;
 
-        Task ExtractAllAttachments(
-            string inputFile,
-            MediaSourceInfo mediaSource,
-            string outputPath,
-            CancellationToken cancellationToken);
+public interface IAttachmentExtractor
+{
+    /// <summary>
+    /// Gets the path to the attachment file.
+    /// </summary>
+    /// <param name="item">The <see cref="BaseItem"/>.</param>
+    /// <param name="mediaSourceId">The media source id.</param>
+    /// <param name="attachmentStreamIndex">The attachment index.</param>
+    /// <param name="cancellationToken">The cancellation token.</param>
+    /// <returns>The async task.</returns>
+    Task<(MediaAttachment Attachment, Stream Stream)> GetAttachment(
+        BaseItem item,
+        string mediaSourceId,
+        int attachmentStreamIndex,
+        CancellationToken cancellationToken);
 
-        Task ExtractAllAttachmentsExternal(
-            string inputArgument,
-            string id,
-            string outputPath,
-            CancellationToken cancellationToken);
-    }
+    /// <summary>
+    /// Gets the path to the attachment file.
+    /// </summary>
+    /// <param name="inputFile">The input file path.</param>
+    /// <param name="mediaSource">The <see cref="MediaSourceInfo" /> source id.</param>
+    /// <param name="cancellationToken">The cancellation token.</param>
+    /// <returns>The async task.</returns>
+    Task ExtractAllAttachments(
+        string inputFile,
+        MediaSourceInfo mediaSource,
+        CancellationToken cancellationToken);
 }