Contributions

Be it a small correction in the documentation or a major new feature, all contributions are welcome. Please, make sure to contact us before working on a contribution so we can agree on the scope and the changes.

General guidelines

Please make sure:

  • your code is clean and efficient

  • you use python typing

  • you test any new functionality

Codebase setup

Install the project locally with:

pip install -e ".[test]"

This will install the project together with the requirements to run testing.

Linting

We use pre-commit to keep a consistent coding style. To run it:

pre-commit run --all-files

This will automatically format the code, except from the flake8 issues which need to be manually resolved. Once fixed, add again your changes with git.

Testing

To run all tests:

DATA_PATH_HNNE=<path to test datasets> pytest

Note the DATA_PATH_HNNE environment variable. This is a location where some torchvision datasets will be stored to be used in some of the tests. Those are currently the test parts of MNIST, FMNIST and CIFAR10 which occupy approximately 200 MB of disk space.

Upload new version to PyPi (for admins)

To upload a version you need to have twine installed and access to the PiPy project: https://pypi.org/project/hnne. To upload a new version:

  • make sure you updated the version of the package in setup.py to the new one

  • also update the version in docs/source/conf.py for consistent documentation

  • ensure the code is merged in the main branch

  • create and push a new tag with the version number, e.g. for version 0.1.5:

git tag 0.1.5
git push origin --tags

  • Create a source distribution:

python setup.py sdist

  • Upload the distribution with twine (you will be asked for your PiPy credentials):

twine upload dist/*