Merge pull request #945 from mitchcapper/archive_extension_hinting_pr

Extension hinting for ReaderFactory for better first try factory success

src/SharpCompress/Factories/TarFactory.cs

@@ -1,10 +1,13 @@
+using System;
 using System.Collections.Generic;
 using System.IO;
+using System.Linq;
 using SharpCompress.Archives;
 using SharpCompress.Archives.Tar;
 using SharpCompress.Common;
 using SharpCompress.Compressors;
 using SharpCompress.Compressors.BZip2;
+using SharpCompress.Compressors.Deflate;
 using SharpCompress.Compressors.LZMA;
 using SharpCompress.Compressors.Lzw;
 using SharpCompress.Compressors.Xz;
@@ -14,6 +17,7 @@ using SharpCompress.Readers;
 using SharpCompress.Readers.Tar;
 using SharpCompress.Writers;
 using SharpCompress.Writers.Tar;
+using GZipArchive = SharpCompress.Archives.GZip.GZipArchive;
 
 namespace SharpCompress.Factories;
 
@@ -39,32 +43,13 @@ public class TarFactory
     /// <inheritdoc/>
     public override IEnumerable<string> GetSupportedExtensions()
     {
-        // from https://en.wikipedia.org/wiki/Tar_(computing)#Suffixes_for_compressed_files
-
-        yield return "tar";
-
-        // gzip
-        yield return "taz";
-        yield return "tgz";
-
-        // bzip2
-        yield return "tb2";
-        yield return "tbz";
-        yield return "tbz2";
-        yield return "tz2";
-
-        // lzma
-        // yield return "tlz"; // unsupported
-
-        // xz
-        // yield return "txz"; // unsupported
-
-        // compress
-        yield return "tZ";
-        yield return "taZ";
-
-        // zstd
-        yield return "tzst";
+        foreach (var testOption in compressionOptions)
+        {
+            foreach (var ext in testOption.KnownExtensions)
+            {
+                yield return ext;
+            }
+        }
     }
 
     /// <inheritdoc/>
@@ -102,6 +87,77 @@ public class TarFactory
 
     #region IReaderFactory
 
+
+    protected class TestOption
+    {
+        public readonly CompressionType Type;
+        public readonly Func<Stream, bool> CanHandle;
+        public readonly bool WrapInSharpCompressStream;
+
+        public readonly Func<Stream, Stream> CreateStream;
+
+        public readonly IEnumerable<string> KnownExtensions;
+
+        public TestOption(
+            CompressionType Type,
+            Func<Stream, bool> CanHandle,
+            Func<Stream, Stream> CreateStream,
+            IEnumerable<string> KnownExtensions,
+            bool WrapInSharpCompressStream = true
+        )
+        {
+            this.Type = Type;
+            this.CanHandle = CanHandle;
+            this.WrapInSharpCompressStream = WrapInSharpCompressStream;
+            this.CreateStream = CreateStream;
+            this.KnownExtensions = KnownExtensions;
+        }
+    }
+
+    // https://en.wikipedia.org/wiki/Tar_(computing)#Suffixes_for_compressed_files
+    protected TestOption[] compressionOptions =
+    [
+        new(CompressionType.None, (stream) => true, (stream) => stream, ["tar"], false), // We always do a test for IsTarFile later
+        new(
+            CompressionType.BZip2,
+            BZip2Stream.IsBZip2,
+            (stream) => new BZip2Stream(stream, CompressionMode.Decompress, false),
+            ["tar.bz2", "tb2", "tbz", "tbz2", "tz2"]
+        ),
+        new(
+            CompressionType.GZip,
+            GZipArchive.IsGZipFile,
+            (stream) => new GZipStream(stream, CompressionMode.Decompress),
+            ["tar.gz", "taz", "tgz"]
+        ),
+        new(
+            CompressionType.ZStandard,
+            ZStandardStream.IsZStandard,
+            (stream) => new ZStandardStream(stream),
+            ["tar.zst", "tar.zstd", "tzst", "tzstd"]
+        ),
+        new(
+            CompressionType.LZip,
+            LZipStream.IsLZipFile,
+            (stream) => new LZipStream(stream, CompressionMode.Decompress),
+            ["tar.lz"]
+        ),
+        new(
+            CompressionType.Xz,
+            XZStream.IsXZStream,
+            (stream) => new XZStream(stream),
+            ["tar.xz", "txz"],
+            false
+        ),
+        new(
+            CompressionType.Lzw,
+            LzwStream.IsLzwStream,
+            (stream) => new LzwStream(stream),
+            ["tar.Z", "tZ", "taZ"],
+            false
+        ),
+    ];
+
     /// <inheritdoc/>
     internal override bool TryOpenReader(
         SharpCompressStream rewindableStream,
@@ -111,84 +167,35 @@ public class TarFactory
     {
         reader = null;
         long pos = ((IStreamStack)rewindableStream).GetPosition();
-
-        if (TarArchive.IsTarFile(rewindableStream))
+        TestOption? testedOption = null;
+        if (!string.IsNullOrWhiteSpace(options.ExtensionHint))
         {
-            ((IStreamStack)rewindableStream).StackSeek(pos);
-            reader = OpenReader(rewindableStream, options);
-            return true;
-        }
-
-        ((IStreamStack)rewindableStream).StackSeek(pos);
-        if (BZip2Stream.IsBZip2(rewindableStream))
-        {
-            ((IStreamStack)rewindableStream).StackSeek(pos);
-            var testStream = new BZip2Stream(
-                SharpCompressStream.Create(rewindableStream, leaveOpen: true),
-                CompressionMode.Decompress,
-                false
+            testedOption = compressionOptions.FirstOrDefault(a =>
+                a.KnownExtensions.Contains(
+                    options.ExtensionHint,
+                    StringComparer.CurrentCultureIgnoreCase
+                )
             );
-            if (TarArchive.IsTarFile(testStream))
+            if (testedOption != null)
             {
-                ((IStreamStack)rewindableStream).StackSeek(pos);
-                reader = new TarReader(rewindableStream, options, CompressionType.BZip2);
-                return true;
+                reader = TryOption(rewindableStream, options, pos, testedOption);
+                if (reader != null)
+                {
+                    return true;
+                }
             }
         }
 
-        ((IStreamStack)rewindableStream).StackSeek(pos);
-        if (ZStandardStream.IsZStandard(rewindableStream))
+        foreach (var testOption in compressionOptions)
         {
-            ((IStreamStack)rewindableStream).StackSeek(pos);
-            var testStream = new ZStandardStream(
-                SharpCompressStream.Create(rewindableStream, leaveOpen: true)
-            );
-            if (TarArchive.IsTarFile(testStream))
+            if (testedOption == testOption)
             {
-                ((IStreamStack)rewindableStream).StackSeek(pos);
-                reader = new TarReader(rewindableStream, options, CompressionType.ZStandard);
-                return true;
+                continue; // Already tested above
             }
-        }
-
-        ((IStreamStack)rewindableStream).StackSeek(pos);
-        if (LZipStream.IsLZipFile(rewindableStream))
-        {
             ((IStreamStack)rewindableStream).StackSeek(pos);
-            var testStream = new LZipStream(
-                SharpCompressStream.Create(rewindableStream, leaveOpen: true),
-                CompressionMode.Decompress
-            );
-            if (TarArchive.IsTarFile(testStream))
+            reader = TryOption(rewindableStream, options, pos, testOption);
+            if (reader != null)
             {
-                ((IStreamStack)rewindableStream).StackSeek(pos);
-                reader = new TarReader(rewindableStream, options, CompressionType.LZip);
-                return true;
-            }
-        }
-
-        ((IStreamStack)rewindableStream).StackSeek(pos);
-        if (XZStream.IsXZStream(rewindableStream))
-        {
-            ((IStreamStack)rewindableStream).StackSeek(pos);
-            var testStream = new XZStream(rewindableStream);
-            if (TarArchive.IsTarFile(testStream))
-            {
-                ((IStreamStack)rewindableStream).StackSeek(pos);
-                reader = new TarReader(rewindableStream, options, CompressionType.Xz);
-                return true;
-            }
-        }
-
-        ((IStreamStack)rewindableStream).StackSeek(pos);
-        if (LzwStream.IsLzwStream(rewindableStream))
-        {
-            var testStream = new LzwStream(rewindableStream);
-            ((IStreamStack)rewindableStream).StackSeek(pos);
-            if (TarArchive.IsTarFile(testStream))
-            {
-                ((IStreamStack)rewindableStream).StackSeek(pos);
-                reader = new TarReader(rewindableStream, options, CompressionType.Lzw);
                 return true;
             }
         }
@@ -196,6 +203,33 @@ public class TarFactory
         return false;
     }
 
+    private static IReader? TryOption(
+        SharpCompressStream rewindableStream,
+        ReaderOptions options,
+        long pos,
+        TestOption testOption
+    )
+    {
+        if (testOption.CanHandle(rewindableStream))
+        {
+            ((IStreamStack)rewindableStream).StackSeek(pos);
+            var inStream = rewindableStream;
+            if (testOption.WrapInSharpCompressStream)
+            {
+                inStream = SharpCompressStream.Create(rewindableStream, leaveOpen: true);
+            }
+            var testStream = testOption.CreateStream(rewindableStream);
+
+            if (TarArchive.IsTarFile(testStream))
+            {
+                ((IStreamStack)rewindableStream).StackSeek(pos);
+                return new TarReader(rewindableStream, options, testOption.Type);
+            }
+        }
+
+        return null;
+    }
+
     /// <inheritdoc/>
     public IReader OpenReader(Stream stream, ReaderOptions? options) =>
         TarReader.Open(stream, options);

src/SharpCompress/Readers/ReaderFactory.cs

@@ -1,6 +1,8 @@
+using System;
 using System.IO;
 using System.Linq;
 using SharpCompress.Common;
+using SharpCompress.Factories;
 using SharpCompress.IO;
 
 namespace SharpCompress.Readers;
@@ -22,8 +24,32 @@ public static class ReaderFactory
 
         long pos = ((IStreamStack)bStream).GetPosition();
 
-        foreach (var factory in Factories.Factory.Factories.OfType<Factories.Factory>())
+        var factories = Factories.Factory.Factories.OfType<Factories.Factory>();
+
+        Factory? testedFactory = null;
+
+        if (!string.IsNullOrWhiteSpace(options.ExtensionHint))
         {
+            testedFactory = factories.FirstOrDefault(a =>
+                a.GetSupportedExtensions()
+                    .Contains(options.ExtensionHint, StringComparer.CurrentCultureIgnoreCase)
+            );
+            if (
+                testedFactory?.TryOpenReader(bStream, options, out var reader) == true
+                && reader != null
+            )
+            {
+                return reader;
+            }
+            ((IStreamStack)bStream).StackSeek(pos);
+        }
+
+        foreach (var factory in factories)
+        {
+            if (testedFactory == factory)
+            {
+                continue; // Already tested above
+            }
             ((IStreamStack)bStream).StackSeek(pos);
             if (factory.TryOpenReader(bStream, options, out var reader) && reader != null)
             {

src/SharpCompress/Readers/ReaderOptions.cs

@@ -16,4 +16,9 @@ public class ReaderOptions : OptionsBase
     public bool DisableCheckIncomplete { get; set; }
 
     public int BufferSize { get; set; } = DefaultBufferSize;
+
+    /// <summary>
+    /// Provide a hint for the extension of the archive being read, can speed up finding the correct decoder.  Should be without the leading period in the form like: tar.gz or zip
+    /// </summary>
+    public string? ExtensionHint { get; set; }
 }