SevenZip

The SevenZip library for .NET provides classes to create, modify and extract different kinds of archives by using the native 7z library from Igor Pavlov. Additionally, it contains the original SDK code which allows compressing and decompressing binary data using the LZMA method.

The library is written in C# and supports .NET 4.8 and .NET6 on Windows, and .NET 8 on Windows and Linux.

License information

MIT

The following binaries embedded in the NuGet package are licensed under the GNU LGPL, GNU LGPL with 'unRAR license restrictions' and BSD 3-clause license:

• 7z.dll
• 7z.so

The files located in src/SevenZip/CoreSdk where published under public domain.

All other files are licensed with the MIT license.

References

• Original SDK code for LZMA compression.
• The UUIDs for the COM interfaces were taken from the 7-Zip C/C++ library.
• The format detection is inspired by the original SevenZipSharp library, which has been archived, but has some forks which are still being maintained.
  • Most signatures can be obtained from Wikipedia - List of file signatures.

Limitations

• Multi-part archives are not supported. If there is a requirement, this may be added in a future release.
• Creation of executable archives is not supported.
• Encryption is not yet fully supported. It can however be set by adding the parameters "he" and/or "em" (for Zip archives only) to the CompressProperties class.

Linux Support

The .NET 8 builds of SevenZip can run under Linux. The native 7z Linux library was build from the original repository at https://github.com/mcmilk/7-Zip by using the following commands:

git clone https://github.com/mcmilk/7-Zip.git
cd 7-zip/CPP/7zip/Bundles/Format7zF
make -f makefile.gcc

The different bundles and the formats they support are explained here: https://github.com/mcmilk/7-Zip/blob/master/DOC/readme.txt.

Referencing the library

Project reference

Clone the project into any location and add a reference to it.

<ItemGroup>
  <ProjectReference Include="..\..\src\SevenZip\SevenZip.csproj" />
</ItemGroup>

The native 7z library is part of this repository, but not linked directly by the SevenZip project. This can be done by importing the Native7zReferences.target file in a .csproj, or with the following snippet:

  <ItemGroup>
    <None Include="..\..\runtimes\win\native\7z.dll" Link="7z.dll" CopyToOutputDirectory="PreserveNewest" />
    <None Include="..\..\runtimes\linux\native\7z.so" Link="7z.so" CopyToOutputDirectory="PreserveNewest" />
  </ItemGroup>

The path to the native library can however be specified by setting the ArchiveConfig.NativeLibraryPath property, which is then passed to an ArchiveReader or ArchiveWriter.

NuGet reference

Not available yet

<ItemGroup>
  <PackageReference Include="Exocad.SevenZip" Version="1.0.0" />
</ItemGroup>

The NuGet package includes the native libraries, a separate reference is not required.

Usage

Extracting an archive

The ArchiveReader can extract specific files only or an entire archive. It operates with System.IO.Stream, which allows extracting to memory, files or anything else that can be represented as stream.

A stream to use for extraction is obtained by invoking a callback passed to the Extract method, which has to return the stream to use for a given entry.

The easiest way to extract an entire archive is to use the ExtractAll method:

using SevenZip

using var reader = new ArchiveReader("archive.7z");

reader.ExtractAll(targetDir, ArchiveFlags.ApplyArchiveEntryTimestampsToFileStreams);

To extract a single entry (or specific entries), the Extract methods can be used with the indices of the entries to extract. The index of an entry can be found by iterating over the ArchiveReader.Entries property.

The following snippet shows how to extract a set of entries:

using SevenZip

bool CanExtract(ArchiveEntry entry)
{
  return ...
}

using var reader = new ArchiveReader("archive.7z");

var indices = reader.Entries
  .Where(entry => CanExtract(entry))
  .Select(entry => entry.Index)
  .TolArray();

reader.Extract(
  indices,
  entry => File.Open(Path.Join(baseDir, entry.Path), FileMode.Create, FileAccess.Write),
  ArchiveFlags.ApplyArchiveEntryTimestampsToFileStreams |
    ArchiveFlags.CloseArchiveEntryStreamAfterExtraction);

Creating or modifying an archive

The ArchiveWriter class can be used to create or update an archive. If the class is initialized with a contructor expecting an ArchiveFormat parameter, a new archive is created - even if an archive already exists at the given location!

To update an existing archive, choose a constructor not expecting the ArchiveFormat. The ArchiveWriter will then detect the format from the existing archive (or throws an exception if it can't).

When adding new files to an archive an archive-path must be specified indicating the full path the new entry should have within the archive.

using SevenZip;
using SevenZip.Extensions;

// Create a new archive (or override an existing one)
using var writer = new ArchiveWriter(ArchiveFormat.SevenZip, "archive.7z");

// Add "./directory/filename.txt" as "archive/filename.txt" to the archive.
writer.AddFile("archive/filename.txt", "./directory/filename.txt");

// Add all files from "otherdir" to the archive, using a specific naming strategy.
writer.AddDirectory("./otherdir", NamingStrategy.RelativeToTopDirectoryInclusive);

// Write the changes
writer.Compress();

To modify or remove existing entries, use the ArchiveWriter.ExistingEntries property to search for the index of the entry to update or delete. ArchiveWriter.ReplaceEntry and ArchiveWriter.DeleteEntry can then be used to prepare the changes.

To actually write the changes, call ArchiveWriter.Compress.

using SevenZip;
using SevenZip.Extensions;

// Create a new archive (or override an existing one)
using var writer = new ArchiveWriter("archive.7z");

// Remove entry with index 10.
writer.RemoveEntry(10);

// Replace entry 11 with the contents of another file.
writer.ReplaceEntry(11, "./contents/anotherfile.txt");

// Write the changes
writer.Compress();

Important
If an existing archive is modified, a local copy of the original data is stored in a temporary file which will be deleted again afterwards when calling Compress. This is required since simultaneous read and write operations are not supported.

Once Compress has been called, the ExistingEntries properties is being updated. In most scenarios, the method should only be called once and the ArchiveWriter should then be disposed again. However, for streams supporting both read and write operations, it may be called multiple times.