Create a GitHub Actions CI workflow for a Python Package

Hi all,

In this post I’m going to show you how to create a GitHub Actions CI workflow that runs on a Python package project and:

Triggers on every sucesfull commit in the main branch
Run the tests and generates the code coverage report
Generate the package and makes it available through an artifact

You can check the result in my GitHub repository

As the source package, I’ve created a very simple Python project for a calculator class, (hey, I know, It’s not a really awesome calculator project, in fact, it’s a really poor one , but it shows an example of how to create and organize a python package project folder and how to set up unit tests and generate code coverage report for it, so I’m quite happy with it as I see it as a good starting point for any new project :) )

So, these were the steps I did to get the workflow up and running:

Creating the template file

First of all I had to create the GitHub Actions template, so I needed to create a folder path like “.github/workflows” inside the project root directory, and after that I created a YAML file inside that directory with the workflow template.

After creating the folder, I just created an empty YAML file called build.pipeline.yml

To start adding content to the template, I used this example in the GitHub documentation page which I found really inspiring.

Firing the workflow automatically

To fire the workflow with every commit in the (main/master) branch I just had to specify the “on” event in the workflow template, defining how I wanted the workflow to be triggered:

create-a-github-actions-ci-pipeline-for-a-python-package-4c4c02dc5f2e

This way I told GitHub that I wanted to run the worklow not only with every commit on (master/main) branch but with every pull request on this branch too.

Running tests and generating the coverage report

Installing dependencies

This project uses unittest as the testing framework and coverage.py for generating code coverage results. As unittest ships with python, I only needed to install coverage.py in the agent before running the tests.

It is also a good recommendation to update pip before any package installation so upgrading pip was also in my list.

And finally, I wanted my workflow to be ready If I add a requirements.txt file in the near future, so I added a condition in the step to install the libraries in the requirements.txt file if necessary

Running tests and coverage

After installing coverage.py I was ready to run the tests, so I added a new script in my workflow template that executes following steps:

go to src directory (that’s where my source code is located inside my repository)
run unit tests
run coverage based on unittests results
generate coverage report and show it on the screen

5. generate coverage report in html format

Publishing code coverage result as a workflow artifact

The last line of the previous script generated an Html version of the code coverage report, so I wanted to download it when the workflow ends, to do so I used the upload artifact action:

I know that when you run the command “coverage html”, a htmlcov folder gets created with all the coverage report formatted as html, so I just needed to archive that folder and upload it as an artifact, which the actions does it at once (lucky me! :) )

Generating the package

Ok! this is the last step of the workflow, as I followed the official python documentation about package creation I knew that for packaging my project I just needed to:

Install PyPA’s build
Execute build

So that’s what I did! after executing this step I just added another one to create and upload the package output as a workflow artifact, same as I did with the code coverage report:

After all these steps, I just uploaded my template to my GitHub repo and I checked out the results, and after some adjustments (nothing works at first try, huh?) I got the results I expected :)

Here you have my entire workflow template:

# GitHub CI build pipeline
name: Awesome Calculator CI build

on:
  push:
    branches:
      - master
      - main
  pull_request:
    branches:
      - master
      - main
jobs:
  build:
    runs-on: ubuntu-latest
    steps:
    - uses: actions/checkout@v2
    - name: Set up python 3.7
      uses: actions/setup-python@v2
      with:
        python-version: 3.7
    - name: Install dependencies
      run: |
        cd src
        python -m pip install --upgrade pip
        pip install coverage
        if  [ -f requirements.txt ]; then pip install -r requirements.txt; fi
    - name: Run Unit Tests and generate coverage report
      run: |
        cd src
        python -m unittest -v
        coverage run -m unittest
        coverage report
        coverage html
    - name: Archive code coverage html report
      uses: actions/upload-artifact@v2
      with:
        name: code-coverage-report
        path: src/htmlcov
    - name: Run package creation
      run: |
        cd src
        python -m pip install --user --upgrade build
        python -m build
    - name: Archive package
      uses: actions/upload-artifact@v2
      with:
        name: awesome-calculator
        path: src/dist

Create a GitHub Actions CI workflow for a Python Package

Create a GitHub Actions CI workflow for a Python Package

Creating the template file

Firing the workflow automatically

Running tests and generating the coverage report

Installing dependencies

Running tests and coverage

Publishing code coverage result as a workflow artifact

Generating the package

Recommend

GitHub - actions/workflow-parser: The parser for GitHub Actions Workflow files

Testing Azure AD-protected APIs, part 4: GitHub Actions CI workflow

解决包含 GitHub Actions Workflow 的分支无法推送的问题

Workflow syntax for GitHub Actions - GitHub Docs

Unpacking Bash shell tips from a GitHub Actions workflow

CI/CD using GitHub Actions Workflow directly from Visual Studio

GitHub Actions: Build Your First Workflow

Lab 9 - GitHub Actions Workflow

7 advanced workflow automation features with GitHub Actions

Maghic: a Swiss army knife for GitHub Actions Workflow

About Joyk