Briefcase Linux AppImage Template

A Cookiecutter template for building Python apps that will run under Linux, packaged as an AppImage.

Using this template

The easiest way to use this project is to not use it at all - at least, not directly. Briefcase is a tool that uses this template, rolling it out using data extracted from a pyproject.toml configuration file.

However, if you do want use this template directly...

1. Install cookiecutter. This is a tool used to bootstrap complex project templates:

   $ pip install cookiecutter
   

2. Run cookiecutter on the template:

   $ cookiecutter https://github.com/beeware/briefcase-linux-appimage-template
   

   This will ask you for a number of details of your application, including the name of your application (which should be a valid PyPI identifier), and the Formal Name of your application (the full name you use to describe your app). The remainder of these instructions will assume a name of my-project, and a formal name of My Project.

3. Obtain a Python Linux support package for x86_64, and extract it into the My Project/My Project.AppDir/usr directory generated by the template. This will give you a self-contained Python install. If installed correctly, there should be a My Project/My Project.AppDir/usr/bin/python3 binary, as well as some other Python-related files.

   Alternatively, you can download the `Python-Linux-support`_ project, and build your own versions of these frameworks. You will need to do this if you need to build for an architecture other than x86_64.

4. Add your code to the template, into the My Project/My Project.AppDir/usr/app directory. At the very minimum, you need to have an app/<app name>/__main__.py file that defines an entry point that will start your application.

   If your code has any dependencies, they should be installed into the My Project/My Project.AppDir/usr/app_packages directory.

If you've done this correctly, a project with a formal name of My Project, with an app name of my-project should have a directory structure that looks something like:

My Project/
    My Project.AppDir/
        usr/
            app/
                my_project/
                    __init__.py
                    __main__.py
                    app.py
            app_packages/
                ...
            bin/
                python3
                ...
            lib/
                ...
            share/
                ...
        com.example.my-project.desktop
    briefcase.toml

This directory can then be compiled into an AppImage using linuxdeploy. Download the linuxdeploy AppImage, and make the binary executable:

$ chmod +x linuxdeploy-x86_64.AppImage

Then compile your AppDir directory (substituting your release version number):

$ VERSION=1.2.3 ./linuxdeploy-x86_64.AppImage --appdir=My\ Project/My\ Project.AppDir -o appimage -d My\ Project/My\ Project.AppDir/com.example.my-project.desktop

This will produce My Project-1.2.3-x86_64.AppImage. This image can given to any other Linux user, and should run without installing any other dependencies.

Next steps

Of course, running Python code isn't very interesting by itself.

To do something interesting, you'll need to work with the native system libraries to draw widgets and respond to user input. The GTK+ GUI library provides Python bindings that you can use to build a user interface. Alternatively, you could use a cross-platform widget toolkit that supports Windows (such as Toga) to provide a GUI for your application.

If you have any external library dependencies (like Toga, or anything other third-party library), you should install the library code into the app_packages directory. This directory is the same as a site_packages directory on a desktop Python install.