README-dev.md

pickpack for devs

If you are thinking of cloning or modifying pickpack, here are some tips, divided into the following sections:

• poetry
  • Installing poetry
  • Installing dependencies
  • Updating dependencies
• ci
• Distributing on PyPi
  • twine
• Testing

poetry

pickpack has been packaged using poetry

Installing poetry

You can either install poetry through your distribution's package manager or with

curl -sSL https://install.python-poetry.org | python3 -

Beware: poetry lock files are not always backwards compatible. Current lock files (poetry 1.4) are not compatible with poetry 1.2.

Installing dependencies

Once you have poetry, you can run the following command to install dependencies:

poetry install

If you want to also install dependencies from a specific group (for example, dev), use:

poetry install --with dev

Updating dependencies

When adding new dependencies or updating existing ones in your pyproject.toml, remember to update the lock file with

poetry lock

If you want to avoid upgrading dependencies, you may run

poetry lock --no-update

ci

If you want to continue using the CI pipeline under .github/workflows/ci.yml, you will need to take a few elements into account:

1. Python version: what versions will your fork support?
2. OSs, especially Ubuntu LTS: what versions will you test with?
   • From experience, testing with a very old version of Ubuntu may lead to fails simply because the image has become unavailable
3. poetry version: as previously said, poetry lock files are not always backwards compatible. Make sure you are running poetry commands with a compatible version of the package.

Distributing on PyPi

There are different ways to distribute packages on PyPi. For pickpack, twine has been used. This one is up to you.

To bundle your package and get it ready for upload:

poetry build

twine

If you choose to use twine, you can install it with

pip install twine

Then, set up your credentials in `$HOME/.pypirc':

[pypi]
	username = __token__
	password = <token generated on PyPi>

Finally, upload your package with

twine upload -r pypi path/to/package/dist/*

Testing

You can either write a script which runs whatever tests you want on its own or alter the tests already specified under tests/test_pickpack.py.

To run the aforementioned file using poetry, use the command:

poetry run python tests/test_pickpack.py