Flit is a simple way to put Python packages and modules on PyPI. It tries to require less thought about packaging and help you avoid common mistakes. See Why use Flit? for more about how it compares to other Python packaging tools.

$ python3 -m pip install flit

Flit requires Python 3 and therefore needs to be installed using the Python 3 version of pip.

Python 2 modules can be distributed using Flit, but need to be importable on Python 3 without errors.

Say you're writing a module foobar — either as a single file foobar.py, or as a directory — and you want to distribute it.

1. Make sure that foobar's docstring starts with a one-line summary of what the module is, and that it has a __version__:

   """An amazing sample package!"""
   
   __version__ = '0.1'

2. Install flit if you don't already have it:

   python3 -m pip install flit
   

3. Run flit init in the directory containing the module to create a pyproject.toml file. It will look something like this:

   [build-system]
   requires = ["flit_core >=2,<4"]
   build-backend = "flit_core.buildapi"
   
   [tool.flit.metadata]
   module = "foobar"
   author = "Sir Robin"
   author-email = "[email protected]"
   home-page = "https://github.com/sirrobin/foobar"

   You can edit this file to add other metadata, for example to set up command line scripts. See the pyproject.toml page of the documentation.

   If you have already got a flit.ini file to use with older versions of Flit, convert it to pyproject.toml by running python3 -m flit.tomlify.

4. Run this command to upload your code to PyPI:

   flit publish
   

Once your package is published, people can install it using pip just like any other package. In most cases, pip will download a 'wheel' package, a standard format it knows how to install. If you specifically ask pip to install an 'sdist' package, it will install and use Flit in a temporary environment.

To install a package locally for development, run:

flit install [--symlink] [--python path/to/python]

Flit packages a single importable module or package at a time, using the import name as the name on PyPI. All subpackages and data files within a package are included automatically.