Merge pull request #399 from drewnoakes/port-unknown-tag-handler

Port UnknownTagHandler
drewnoakes · Feb 5, 2024 · 5f22ae3 · 5f22ae3
2 parents 60750c7 + 1f89a2c
commit 5f22ae3
Show file tree

Hide file tree

Showing 3 changed files with 71 additions and 9 deletions.
diff --git a/MetadataExtractor.Tools.FileProcessor/DeconstructionExtensions.cs b/MetadataExtractor.Tools.FileProcessor/DeconstructionExtensions.cs
@@ -0,0 +1,17 @@
+// Copyright (c) Drew Noakes and contributors. All Rights Reserved. Licensed under the Apache License, Version 2.0. See LICENSE in the project root for license information.
+
+#if NETFRAMEWORK
+
+namespace System.Collections.Generic;
+
+internal static class DeconstructionExtensions
+{
+    [DebuggerStepThrough]
+    public static void Deconstruct<TKey, TValue>(this KeyValuePair<TKey, TValue> pair, out TKey key, out TValue value)
+    {
+        key = pair.Key;
+        value = pair.Value;
+    }
+}
+
+#endif
diff --git a/MetadataExtractor.Tools.FileProcessor/Program.cs b/MetadataExtractor.Tools.FileProcessor/Program.cs
@@ -14,8 +14,6 @@
 
 namespace MetadataExtractor.Tools.FileProcessor
 {
-    // TODO port UnknownTagHandler
-
     internal static class Program
     {
         private static int Main(string[] args)
@@ -191,11 +189,11 @@ private static int ProcessRecursively(string[] args)
                     // If "--markdown" is specified, write a summary table in markdown format
                     fileHandler = new MarkdownTableOutputHandler();
                 }
-//                else if (arg == "--unknown")
-//                {
-//                    // If "--unknown" is specified, write CSV tallying unknown tag counts
-//                    fileHandler = new UnknownTagHandler();
-//                }
+                else if (arg == "--unknown")
+                {
+                    // If "--unknown" is specified, write CSV tallying unknown tag counts
+                    fileHandler = new UnknownTagHandler();
+                }
                 else if (arg == "--log-file")
                 {
                     if (i == args.Length - 1)
@@ -224,8 +222,7 @@ private static int ProcessRecursively(string[] args)
                 return 1;
             }
 
-            if (fileHandler is null)
-                fileHandler = new BasicFileHandler();
+            fileHandler ??= new BasicFileHandler();
 
             var stopwatch = Stopwatch.StartNew();
 

diff --git a/MetadataExtractor.Tools.FileProcessor/UnknownTagHandler.cs b/MetadataExtractor.Tools.FileProcessor/UnknownTagHandler.cs
@@ -0,0 +1,48 @@
+// Copyright (c) Drew Noakes and contributors. All Rights Reserved. Licensed under the Apache License, Version 2.0. See LICENSE in the project root for license information.
+
+namespace MetadataExtractor.Tools.FileProcessor;
+
+/// <summary>
+/// Keeps track of unknown tags.
+/// </summary>
+internal sealed class UnknownTagHandler : FileHandlerBase
+{
+    private readonly record struct Key(string DirectoryName, int TagType);
+
+    private readonly Dictionary<Key, int> _occurrenceByKey = [];
+
+    public override void OnExtractionSuccess(string filePath, IList<Directory> directories, string relativePath, TextWriter log, long streamPosition)
+    {
+        base.OnExtractionSuccess(filePath, directories, relativePath, log, streamPosition);
+
+        foreach (Directory directory in directories)
+        {
+            foreach (Tag tag in directory.Tags)
+            {
+                // Only interested in unknown tags (those without names)
+                if (tag.HasName)
+                    continue;
+
+                Key key = new(directory.Name, tag.Type);
+
+                _occurrenceByKey.TryGetValue(key, out int count);
+                _occurrenceByKey[key] = count + 1;
+            }
+        }
+    }
+
+    public override void OnScanCompleted(TextWriter log)
+    {
+        base.OnScanCompleted(log);
+
+        var results = _occurrenceByKey
+            .OrderByDescending(pair => pair.Value)
+            .ThenBy(pair => pair.Key.DirectoryName)
+            .ThenBy(pair => pair.Key.TagType);
+
+        foreach ((Key key, int count) in results)
+        {
+            log.WriteLine($"{key.DirectoryName}, 0x{key.TagType:X4}, {count}");
+        }
+    }
+}