Intended Use

req2flatpak is for programmers who want to package a python application using flatpak.

The req2flatpak script takes python package requirements as input, e.g., as requirements.txt file. It allows to specify the target platform’s python version and architecture. The script outputs an automatically generated flatpak-builder build module. The build module, if included into a flatpak-builder build manifest, will install the python packages using pip.

Getting Started

Install req2flatpak using pip (or any other python package installer), e.g., by running pip install req2flatpak.

Prepare a requirements.txt file with the packages that you want to install as part of a flatpak build. You need to specify exact versions for all the packages including their dependencies. (Various tools exist to help creating such a requirements file with fully resolved dependencies and frozen package versions; for example, you can use pip-compile or poetry export to export a suitable requirements.txt file).

Run req2flatpak. For example, run this command to generate a flatpak-builder build module from your requirements.txt file:

./req2flatpak.py --requirements-file requirements.txt --target-platforms 310-x86_64 310-aarch64

When invoked like this, req2flatpak will read the requirements file, query pypi about available downloads for the requirements, choose appropriate downloads for the specified target platforms, and generate a flatpak-builder build module. The module, if included in a flatpak-builder build manifest, will install the required packages using pip. The commandline option to define target platforms uses the format <pythonversion>-<architecture>. To learn more about available commandline options, run req2flatpak.py --help.

Note that programmatic usage of req2flatpak is also possible. This means you can write a python script to invoke req2flatpak, which allows to tweak the desired behavior in many ways.

Go read the documentation to learn more about req2flatpak’s commandline and Python APIs. The documentation includes further examples to help you get started quickly.

Documentation

See https://johannesjh.github.io/req2flatpak/

Contributing

req2flatpak is developed in an open-source, community-driven way, as a voluntary effort in the authors’ free time.

All contributions are greatly appreciated… pull requests are welcome, and so are bug reports and suggestions for improvement. See req2flatpak’s documentation for how to set up a development environment and how to contribute back to req2flatpak.

Related Work

The flatpak-pip-generator script is very similar to this project. Both scripts basically serve the same purpose, and this project took a lot of inspiration from flatpak-pip-generator. In fact, this project was created when we discussed feature request #296 in flatpak-pip-generator. A prototype followed from this feature request, and since it was written from scratch, the prototype became this separate project.

Comparison between flatpak-pip-generator and req2flatpak.py: Each of the two projects has its own benefits. A comparison will likely change over time. As in Oct, 2022, in my personal opinion (johannesjh), I see the following similarities and differences:

• Both projects generate build modules for flatpak-builder.

• Both projects consist of a single script file with minimal dependencies, and are thus very easy to install.

• flatpak-pip-generator resolves dependencies and freezes dependency versions, whereas req2flatpak.py asks the user to provide a fully resolved list of dependencies with frozen dependency versions. Various tools exist which make this easy, e.g., pip, pip-compile and poetry.

• flatpak-pip-generator is older and thus likely to be more mature. It supports more commandline options and probably has a more complete feature set.

• req2flatpak.py is faster. The script itself runs faster because it does not need to download package files in order to generate the build module. And the flatpak build runs faster because all packages (from the entire requirements.txt file) are installed in a single call to pip install.

• req2flatpak.py re-implements some functionality of pip. In contrast, flatpak-pip-generator uses pip’s official functionality. Specifically, req2flatpak.py re-implements how pip queries available downloads from pypi and how pip chooses suitable downloads to match a given target platform.

• req2flatpak.py prefers binary wheels, whereas flatpak-pip-generator prefers source packages.

License

req2flatpak is MIT-licensed, see the COPYING file.

v0.3.2 (2026-11-15)

Improvements:

• Relaxed the version constraint for the packaging dependency from ^21.3 to >=21.3, see #92

Fixes:

• Close output stream which was causing empty outfile files on Python 3.14, see #91

Full Changelog: https://github.com/johannesjh/req2flatpak/compare/v0.3.1…v0.3.2

v0.3.1 (2025-03-30) “Python 3.12”

Highlight: Compatibility with Python 3.12

Improvements for Development:

• Improved CI #73 #79 #82

• Added devcontainers configuration #65

• Now using ruff for linting #87

Fixes:

• Fixed an error when no urls are found in a pypi release in #71

Maintenance and dependency updates:

• Updated dependencies in #62 #74

• Improved docs in #77 #81 #86

• Now using packaging instead of the deprecated pkg_resources package #76

Full Changelog: https://github.com/johannesjh/req2flatpak/compare/v0.2…v0.3

v0.2 (2023-06-04) “Yaml”

Highlight: req2flatpak can now generate yaml output.

Features:

• Adds yaml output. It is now possible to generate yaml output by specifying a .yaml file extension or by specifying the --yaml commandline option.

Bugfixes:

• Fixes sorting when architecture is None - #34

• Fixes exception handling with invalid platforms - #39

v0.1 (2022-12-23) “Initial Release”

Highlight: This is the initial release of req2flatpak.

Features:

• Generates a flatpak-builder build module for given python package requirements; the module will install the required packages as part of a flatpak build.

• This initial release already comes with documentation, a clean code style, and automated tests that are run using continuous integration.