MozJPEG Lossless Optimization

This library optimizes JPEGs losslessly using MozJPEG.

To reduce the file sizes,

• the Huffman table of the JPEGs is optimized,
• the baseline JPEGs are converted to progressive JPEGs,
• and any metadata and ICC profiles are removed.

The JPEGs optimized with this library are identical to what you get using the jpegtran tool from MozJPEG with the -optimize, -progressive and -copy none options.

Usage

Optimizing (losslessly) a JPEG:

import mozjpeg_lossless_optimization

with open("./image.jpg", "rb") as input_jpeg_file:
    input_jpeg_bytes = input_jpeg_file.read()

output_jpeg_bytes = mozjpeg_lossless_optimization.optimize(input_jpeg_bytes)

with open("./out.jpg", "wb") as output_jpeg_file:
    output_jpeg_file.write(output_jpeg_bytes)

Converting an image to an optimized JPEG (using Pillow):

from io import BytesIO

from PIL import Image  # pip install pillow
import mozjpeg_lossless_optimization


def convert_to_optimized_jpeg(input_path, output_path):
    jpeg_io = BytesIO()

    with Image.open(input_path, "r") as image:
        image.convert("RGB").save(jpeg_io, format="JPEG", quality=90)

    jpeg_io.seek(0)
    jpeg_bytes = jpeg_io.read()

    optimized_jpeg_bytes = mozjpeg_lossless_optimization.optimize(jpeg_bytes)

    with open(output_path, "wb") as output_file:
        output_file.write(optimized_jpeg_bytes)


convert_to_optimized_jpeg("input.png", "optimized.jpg")

Install

From PyPI

To install MozJPEG Lossless Optimization from PyPI, just run the following command:

pip install mozjpeg-lossless-optimization

From Sources

To install MozJPEG Lossless Optimization, MozJPEG will be compiled, so you will need a C compilator and cmake. On Debian / Ubuntu you can install everything you need with the following command:

sudo apt install build-essential cmake python3 python3-dev python3-pip python3-setuptools

Once everything installed, clone this repository:

git clone https://github.com/wanadev/mozjpeg-lossless-optimization.git

Then navigate to the project's folder:

cd mozjpeg-lossless-optimization

Retrieve submodules:

git submodule init
git submodule update

And finally build and install using the following command:

python3 setup.py install

Hacking

Get the source and build C lib and module:

# Install system dependencies
sudo apt install build-essential cmake python3 python3-dev python3-pip python3-setuptools

# Get the sources
git clone https://github.com/wanadev/mozjpeg-lossless-optimization.git
cd mozjpeg-lossless-optimization
git submodule init
git submodule update

# Create and activate a Python virtualenv
python3 -m venv __env__
source __env__/bin/activate

# Install Python dependencies in the virtualenv
pip install cffi

# Build MozJPEG
# This will generate files in ./mozjpeg/build/ folder
python setup.py build

# Build the CFFI module "in-place"
# This will create the ./mozjpeg_lossless_optimization/_mozjpeg_opti.*.so file on Linux
python ./mozjpeg_lossless_optimization/mozjpeg_opti_build.py

Lint:

pip install nox
nox -s lint

Run tests:

pip install nox
pip -s test

Licenses

MozJPEG Lossless Optimization is licensed under the BSD 3 Clause license. See the LICENSE file for more information.

MozJPEG is covered by three compatible BSD-style open source licenses. See its license file for more information.

Changelog

• [NEXT] (changes on master but not released yet):
  • Nothing yet ;)
• v1.1.3:
  • Added Python 3.12 support (@flozz, #6)
  • Removed Python 3.7 support (@flozz)
• v1.1.2:
  • Added Python 3.11 support
• v1.1.1:
  • Fix sdist package: missing MozJPEG source files added
• v1.1.0:
  • Update mozjpeg to latest master commit: fd56921 (v4.1.1+)
• v1.0.2:
  • arm64 and universal2 wheels for macOS on Apple Silicon
  • x86 and x68_64 wheels for musl-based Linux distro (Alpine,...)
• v1.0.1: Python 3.10 support and wheels
• v1.0.0: Handle JPEG decompression errors
• v0.9.0: First public release