libffrs

Fairly Fast & Flexible Reed-Solomon Coding

This library uses a variation Intel’s Slicing-by-8 algorithm for the encoding routines, achieving some hefty speeds depending on the code sizes. See Benchmarks.

The core implementation is done in C++ and exported to Python with pybind11.

Basic usage

import ffrs

message = bytearray(b'Hello World!')

# Create encoder/decoder for 8 parity bytes
rs8 = ffrs.RS256(ecc_len=8)

# Allocate space for correction code
data = message + bytearray(rs8.ecc_len)

# Compute parity bytes
rs8.encode(data)

# Corrupt some bytes (up to ecc_len/2)
data[3] ^= 0xaa
data[9] ^= 0x55

# Restore partially corrupted message
rs8.decode(data)

# Remove parity bytes
recovered_message = data[:-rs8.ecc_len]

assert message == recovered_message

Installation

git clone --recursive https://github.com/zydd/libffrs.git
pip install ./libffrs

Documentation

Auto-generated documentation can be found at:

• https://zydd.github.io/libffrs/

Building the documentation

Documentation for this project is generated automatically using Sphinx based on class/method signatures and doc strings.

cd ./libffrs
pip install 'sphinx>=7.2' 'furo==2023.9.10'
make -C ./docs/ html

Documentation files will be created in ./build/html

License

libffrs is licensed under the Apache License, Version 2.0