ShocoSharp: a fast compressor for short strings

based on shoco - a C implementation by Christian Schramm

ShocoSharp is a .NET library to compress and decompress short strings. The default compression model is optimized for English words, but you can generate your own compression model based on your specific input data.

ShocoSharp is free software, distributed under the MIT license.

Quick Start

on NuGet

Install-Package ShocoSharp

API

For very basic usage the following global helper methods are offered:

byte[] Shoco.Compress(string Data);
string Shoco.DecompressString(byte[] Data);

A range of overloads provide great flexibility and the opportunity to optimize use of the library.

The global helper methods operate on the Shoco.DefaultModel compression model. By default, the ShocoCompatibleEnglishWordsModel is used (to provide compatibilty with other implementations). You can substitute the global default model or use any other model directly.

The library ships with several pre-generated compression models, or you can generate your own model from your specific sample data.

How It Works

This library is ported from its original C implementation. See Christian Schramm's excellent explanation at the shoco repository.

Generating Compression Models

Maybe your typical input isn’t English words. Maybe it’s German or French – or whole sentences. Or file system paths. Or URLs. While the standard compression model of ShocoSharp should work for all of these, it might be worthwhile to train ShocoSharp for this specific type of input data.

Fortunately, that’s really easy: ShocoSharp includes a compression model generator which will train a model on your sample data. This model can be used directly after training, or exported to a C# class or C header file for use in your project.

// generate model
var myModel = ShocoModelGenerator.GenerateModelFromFile(@"C:\MyTrainingData\MyData.txt");

// use the model directly
var compressedString = myModel.Compress("sample string");
var originalString = myModel.Decompress(compressedString);

// write out the model
myModel.WriteAsCSharpClass(Filename: @"C:\MyTrainingData\MyModel.cs", Name: "MyShocoModel")

As with the compression/decompression methods there are many overloads that provide great flexibility when generating models. All features provided by the original python model generator are available.

If you use the python model generator to create your compression model, you can load and convert this into ShocoSharp with ShocoModel.ReadFromCHeader(string Filename).

Credits

ShocoSharp is based on shoco, written by Christian Schramm which was released under the MIT license.