Poetry plugin

The Poetry plugin can be used for Python projects that use the Poetry build system.

Keys

This plugin provides the following unique keys.

poetry-export-extra-args:

Type: list of strings

Extra arguments to pass at the end of the poetry export command.

poetry-pip-extra-args:

Type: list of strings

Extra arguments to pass to pip install when installing dependencies.

poetry-with:

Type: list of strings

Dependency groups to include. By default, only the main dependencies are included.

Environment variables

This plugin also sets environment variables in the build environment. User-set environment variables will override these values. Users may also set environment variables to configure Poetry using the build-environment key.

PARTS_PYTHON_INTERPRETER

Default: python3

Either the interpreter binary to search for in PATH or an absolute path to the interpreter (e.g. ${CRAFT_STAGE}/bin/python).

PARTS_PYTHON_VENV_ARGS

Additional arguments passed to python -m venv.

Dependencies

Whether the Python interpreter needs to be included in the snap depends on its confinement. Specifically:

  • Projects with strict or devmode confinement can safely use the base snap’s interpreter, so they typically do not need to include Python.

  • Projects with classic confinement cannot use the base snap’s interpreter and thus must always bundle it (typically via stage-packages).

  • In both cases, a specific/custom Python installation can always be included in the snap. This can be useful, for example, when using a different Python version or building an interpreter with custom flags.

Snapcraft will prefer an included interpreter over the base’s, even for projects with strict and devmode confinement.

How it works

During the build step, the plugin performs the following actions:

  1. Create a virtual environment directly into the ${CRAFT_PART_INSTALL} directory.

  2. Use poetry export to create a requirements.txt file in the project’s build directory.

  3. Use pip to install the packages referenced in requirements.txt into the virtual environment, without any additional dependencies.

  4. Use pip to install the source package without any additional dependencies.

  5. Run pip check to ensure the virtual environment is consistent.