Automating Tests with GitHub Actions

diagram what a CI run does with a push mermaid

  flowchart TD
  A["git push / pull request"] --> B["GitHub Actions starts a fresh runner"]
  B --> C["checkout at that commit"]
  C --> D["set up Python"]
  D --> E["install from requirements.txt"]
  E --> F{"install succeeded?"}
  F -->|no| G["fail -- a dependency was never pinned"]
  F -->|yes| H["lint / type check"]
  H --> I["pytest"]
  I --> J{"all green?"}
  J -->|no| K["the PR is marked failing"]
  J -->|yes| L["merge is allowed"]
  C -.->|"clean machine"| M["catches 'works on my laptop'"]

The value of CI is not that it runs your tests -- you can do that -- but that it runs them somewhere you did not configure, from a clean checkout, on every push. Most of what it catches is not a broken test; it is a missing dependency, a file that was never committed, or something that only worked because of state on one machine.

What GitHub Actions is

GitHub Actions runs workflows on:

• push
• pull_request
• schedules

Minimal Python test workflow

Create .github/workflows/tests.yml:

yaml

name: tests
 
on:
  pull_request:
  push:
 
jobs:
  test:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
 
      - name: Set up Python
        uses: actions/setup-python@v5
        with:
          python-version: "3.11"
 
      - name: Install dependencies
        run: |
          python -m pip install --upgrade pip
          pip install -r requirements.txt
 
      - name: Run tests
        run: |
          pytest -q

Tips

• cache pip if workflows are slow
• upload coverage/html reports as artifacts
• run lint/type checks as separate jobs