Introduction

Python Package Cookiecutter Code to help make faster libraries. The objective of this repo is to concentrate on developing source code by standardizing the ecosystem of tools for development

Summary

Writing code helps us solve a business problem. However writing good code will make it reusable. Writing good code with tests around it will help convert the code into a python library (or module or package) usable by fellow team members or to the open source community.

A developer will want to go through this process in a seamless manner. This article and repo will define a proven process and provide guidelines to write a python library.

#TODO

• Convert library to a cookiecutter template. Helps parametrize the library name.
• Use GitHub actions to include testing, test coverage etc.

For library guidance on a typical data science, see see for data science projects

Usage

The recommended practices for developing python libraries are given below:

https://realpython.com/pypi-publish-python-package/

Using toml file, bumpver and twine

To see instructions using setup.py & twine, without Version Bump, see using setup.py

CI

More CI/CD streamlining for python packages:

• use cookiecutter to generate a package template
• set up travis CI for auto deployment of package to pypi

#TODO

• Convert library to a cookiecutter template. Helps parametrize the library name.

https://github.com/audreyfeldroy/cookiecutter-pypackage https://cookiecutter-pypackage.readthedocs.io/en/latest/tutorial.html https://pypi.org/project/cookiecutter/

https://github.com/boromir674/cookiecutter-python-package https://github.com/boromir674/cookiecutter-python-package/tox.ini

https://youtu.be/ugGu8fHWFog (A data science project example folder)

Github

Helps cover testing, test coverage, etc. https://docs.github.com/en/actions/automating-builds-and-tests/building-and-testing-python https://hynek.me/articles/python-github-actions/

Travis CI, No Free Support?

https://jacobtomlinson.dev/posts/2021/testing-and-continuous-integration-for-python-packages-with-github-actions/

https://github.com/ksator/continuous-integration-with-python

https://medium.com/swlh/automate-python-testing-with-github-actions-7926b5d8a865

Testing a Package Locally

Using .toml file (Working)

A package can be imported locally from another code and thoroughly tested as well if required. Editable install is the best way to achieve this. The steps to do so are:

• Change to the current working directory where the pyproject.toml file is located
• Execute the following command to install the package locally
  • python -m pip install -e .
• This installation uses the files in the current working directory

Using conda-build (Did not work)

Building A Package Locally

• Add following package to the base environment -Install conda-buiild -Conda install conda-build
• Utilize the below to build the package in current path. A specific path can also be specified.
  • Conda develop .
  • https://docs.conda.io/projects/conda-build/en/latest/user-guide/tutorials/build-pkgs.html

Writing Tests

• Write tests. Preferably utilize pytest.
• Example test and file structure
  • https://github.com/jumptrading/luddite
  • Utilized test_package.py for all tests
  • pytest.ini file for pytest configurations
  • Utilize github test workflows
  • https://github.com/jumptrading/luddite/blob/master/.github/workflows/tests.yml
• https://tox.wiki/en/latest/

References

https://www.freecodecamp.org/news/build-your-first-python-package/

https://python-packaging-tutorial.readthedocs.io/en/latest/setup_py.html

https://packaging.python.org/

Guidelines to contribute to libraries: https://pandas.pydata.org/docs/development/contributing.html#contributing

https://realpython.com/pypi-publish-python-package/

https://github.com/johnthagen/python-blueprint