(this project is no longer maintained, please move to http://github.com/richox/orz)

fast and lightweight compression library and utility.

Libzling is an improved lightweight compression utility and library. libzling uses fast order-1 ROLZ (16MB block size and 10MB dictionary size) followed with Huffman encoding, making it 3 times as fast as gzip on compressing, while still getting much better compression ratio and decompression speed.

Simple benchmark with enwik8(100,000,000 bytes), also on Large Text Compression Benchmark (thanks to Matt Mahoney)

CPU: Intel(R) Xeon(R) CPU E5-2620 v3 @ 2.40GHz
MEM: 65726592 kB
OS:  Linux 2.6.32.43-tlinux-1.0.14-default

You can build and install libzling automatically by cmake with the following command:

cd ./build
cmake -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=/path/to/install
make
make install

Libling provides simple and lightweight interface. here is a simple program showing the basic usage of libzling. (compiled with g++ -Wall -O3 zling_sample.cpp -o zling_sample -lzling)

#include "libzling/libzling.h"

int main() {
    // compress
    {
        const int level = 0;  // valid levels: 0, 1, 2, 3, 4
        FILE* fin = fopen("./1.txt", "rb");
        FILE* fout = fopen("./1.txt.zlng", "wb");

        baidu::zling::FileInputter  inputter(fin);
        baidu::zling::FileOutputter outputter(fout);

        baidu::zling::Encode(&inputter, &outputter, level);
        fclose(fin);
        fclose(fout);
    }

    // decompress
    {
        FILE* fin = fopen("./1.txt.zlng", "rb");
        FILE* fout = fopen("./2.txt", "wb");

        baidu::zling::FileInputter  inputter(fin);
        baidu::zling::FileOutputter outputter(fout);

        baidu::zling::Decode(&inputter, &outputter);
        fclose(fin);
        fclose(fout);
    }
    return 0;
}

However libzling supports more complicated interface, see ./demo/zling.cpp for details.