bmaheshwari / python-project-template

Description

A python template project following modern practices. We will work with a simple stringutils library with three basic functions.

Steps

Create a simple library

Create a simple stringutils library. I included only three functions in it to keep it simple.

Add docstrings

I added function-level and file-level strings (where necessary). For function level strings, I followed the numpy docstring format.

Add requirements

I included requirements i.e. base.txt and check.txt. I added them under the requirements folder.

Write tests

I added some tests for the functions to verify if they are working well. I am using Pytest for that.

pytest --cov stringutils --cov-report=term-missing --tb=line
pytest --cov . --cov-report=xml --tb=line

Note

• We can manually import our library functions for tests using __init__.py.
• Or we can leave it empty and use [setup.py](http://setup.py) to install and use the package in our test.

Add pre-commit hooks

Next, I added pre-commit hooks to make sure that my code is in compliance with the best python practices.

pip install pre-commit
pre-commit install

It will work with each commit. To run manually:

pre-commit run --all-files

GitHub Workflows

We have two workflows i.e. test.yml to test our code and lint.yml to apply pre-commit hooks upon push and pull-requests to the main/master branch. I am also using Codecov action to upload test coverage report. Check the reference link for more details.

Packaging

• Create setup.py.

  • Note: If you face issues with requirements paths, create a MANIFEST.in file like mine.

• Create account on pypi, build and upload.

  • Create ~HOME/pypirc file and add credentials in it:

  [pypi]
    username = __token__
    password = <api-token>
  
  

  • Build the package.

  pip install build
  python -m build
  
  

  • Upload the package.

  twine check dist/*
  twine upload dist/*