python-appimage

AppImage distributions of Python

Ready to use AppImages of Python are available as GitHub releases.

Quickstart

We provide relocatable Python runtimes as AppImages. These runtimes are extracted from manylinux Docker images. The corresponding images are available as GitHub releases. They are labeled according to wheels compatibility tags. Our Python AppImages are updated weekly.

Running Python from these AppImages is as simple as downloading a single file and changing its mode to executable, e.g. as:

wget https://github.com/niess/python-appimage/releases/download/\
python3.8/python3.8.2-cp38-cp38-manylinux1_x86_64.AppImage
chmod +x python3.8.2-cp38-cp38-manylinux1_x86_64.AppImage

./python3.8.2-cp38-cp38-manylinux1_x86_64.AppImage

This should start a Python 3.8 interactive session on almost any Linux provided that fuse is available. Note that on WSL1 since fuse is not supported you will need to extract the AppImage as explained hereafter.

The workflow described previously is enough if you only need vanilla Python with its standard library. However, if you plan to install extra packages we recommend extracting the AppImage, e.g. as:

./python3.8.2-cp38-cp38-manylinux1_x86_64.AppImage --appimage-extract
mv squashfs-root python3.8
rm -f python3.8.2-cp38-cp38-manylinux1_x86_64.AppImage

export PATH="$(pwd)/python3.8/usr/bin:$PATH"

Then, extra packages can be installed to the extracted AppDir using pip. For example upgrading pip can be done as:

pip install --upgrade pip

For applications developers

Python AppImages are built using the python_appimage Python package. You can get it from GitHub or PyPI. Examples of usage can be found by browsing GitHub workflows.

The python_appimage package also allows to build basic Python apps from an existing Python AppImage and a recipe folder. The recipe folder contains the app metadata, a Python requirements file and an entry point script. Examples of recipes can be found on GitHub in the applications folder.

Alternatively, you can also manualy extract one of the Python AppImages as explained above and directly modify the content, e.g. pip install your custom packages. Then, simply rebuild the AppImage using your favourite tool, e.g. appimagetool, linuxdeploy or python-appimage.

Make releases with AppImage in GitHub release assets with Rever

We recommend to test this process in safe place. The best way is to fork or copy the repository you're working on to the distinct GitHub repository.

How to release AppImages with release your app on GitHub:

1. You have PyPi package with setup.py file and pip install -U . works perfect.

2. You have AppImage description files like in applications/xonsh directory. And your AppImage can be build with:

   python -m python_appimage build app ./path/to/package
   

3. Install rever:

   pip install -U rever
   

4. Create appimage directory near setup.py with AppImage description files and rename the appimage/requirements.txt to appimage/pre-requirements.txt. In release process your package directory will be added to pre-requiriments.txt and created requirements.txt.

5. Create rever.xsh file near setup.py:

   $ACTIVITIES = ['tag', 'push_tag', 'appimage', 'ghrelease']
   
   $GITHUB_ORG = 'anki-code'
   $PROJECT = $GITHUB_REPO = 'mypackage'
   
   $TAG_REMOTE = 'git@github.com:anki-code/mypackage.git'
   $TAG_TARGET = 'master'
   $PUSH_TAG_PROTOCOL='ssh'
   
   # The name of your AppImage will be `<Name from .desktop-file>-<system>.AppImage` (`xonsh-x86_64.AppImage` for example).
   $GHRELEASE_ASSETS = ['mypackage-x86_64.AppImage']
   

6. Run check and make release. In this example we have 4 steps: tag, push_tag, appimage, ghrelease. This means that rever will create 0.0.1 tag, push it to the remote, then build AppImage and create GitHub release:

   rever check
   rever 0.0.1
   

7. Check the GitHub release page for AppImage assets

Projects using python-appimage

• grand/python - Contained, portable and modern python for GRAND running from an AppImage
• xxh - Bring your favorite shell wherever you go through the ssh
• xonsh - Python-powered, cross-platform, Unix-gazing shell language and command prompt