Machine Learning/Data Science Repository Template

Components

poetry as a production-ready dependency manager for python. It has any nice features like defining the python version, pinning transitive dependencies or creating a file so all developers use the exact same versions.
python specific .gitignore file
pre-commit hooks allow you do iterate faster by doing code formatting checks locally
ci-cd config for github to automatically execute the tests

(optional): install pyenv to comfortably switch between python versions and install a recent python version
activate the installed python version with $ pyenv local *python_version* or use the already present python version
define the python version in pyproject.toml and in .github/workflows/pytest.yaml and the .pre-commit-config.yaml
update the project name in pyproject.toml
(optional): point poetry to your pyenv python version $ poetry env use ~/.pyenv/versions/*python_version*/bin/python
install dependencies with poetry install. Commit the file poetry.lock that contains the installed dependencies.
activate the environment with poetry shell and run pre-commit install to set up the git hook scripts in the .git directory of your project
run pre-commit autoupdate to update the repository references in the .pre-commit-config.yaml file

When all hooks combined in the pipeline the workflow look like this (image credit)