Introduction

Objectif : Transform a simple pipeline into a robust industrial pipeline, respecting the modern standards of packaging and software engineering.

In the first part of this series, we built a simple but effective CI/CD pipeline to automate testing and publishing a Python package to PyPI.

It is now time to professionalize this pipeline. In this second part, we are going to:

  • Migrate to the modern pyproject.toml standard,

  • Add quality and safety tools (Black, Mypy, Bandit, Safety)

  • Set up multi-version tests,

  • Integrate a gradual deployment via Test PyPI,

  • Automate versioning and improve pipeline monitoring.

Get ready to move from a functional pipeline to a professional CI/CD infrastructure.

Deploying a Python application to PyPI requires a robust CI/CD pipeline that automates testing, building, and publishing packages. This article details the setup of a complete pipeline using GitHub Actions for a Python CLI application, based on the best practices of the modern Python ecosystem.

CI/CD Pipeline Architecture

The pipeline we are going to build follows a multi-step approach:

@startuml
!theme plain
title Python CI/CD pipeline to PyPI
skinparam backgroundColor #f8f9fa
skinparam componentStyle rectangle

rectangle "Developer" as dev
rectangle "GitHub Repository" as repo {
  rectangle ".github/workflows/" as workflows
  rectangle "tests/" as tests
  rectangle "setup.py / pyproject.toml" as setup
  rectangle "requirements.txt" as req
}

rectangle "GitHub Actions" as actions {
  rectangle "Test Runner" as test_runner
  rectangle "Build" as build
  rectangle "Security Scan" as security
  rectangle "(No output, as there is no French text provided to translate.)" as quality
}

rectangle "PyPI" as pypi {
  rectangle "Test PyPI" as test_pypi
  rectangle "Production PyPI" as prod_pypi
}

rectangle "Users" as users

dev --> repo : push/PR
repo --> actions : trigger workflow
actions --> test_runner : run tests
actions --> security : security checks
actions --> quality : code quality
actions --> build : build package
build --> test_pypi : deploy (pre-release)
build --> prod_pypi : deploy (release)
prod_pypi --> users : install package

@enduml

Project Structure

A Python CLI application ready for distribution must adhere to a standardized structure:

playlist-downloader/
├── .github/
│   └── workflows/
│       ├── ci.yml
│       ├── release.yml
│       └── security.yml
├── src/
│   └── playlist_downloader/
│       ├── __init__.py
│       ├── cli.py
│       ├── core/
│       └── adapters/
├── tests/
│   ├── unit/
│   ├── integration/
│   └── conftest.py
├── docs/
├── pyproject.toml
├── requirements.txt
├── requirements-dev.txt
├── MANIFEST.in
├── README.md
├── LICENSE
└── CHANGELOG.md

Package Configuration with pyproject.toml

The file`pyproject.toml`is the modern standard for configuring Python packages:

[build-system]
requires = ["setuptools>=45", "wheel", "setuptools_scm>=6.2"]
build-backend = "setuptools.build_meta"

[project]
name = "playlist-downloader"
authors = [
    {name = "Christophe Hérolivier", email = "[email protected]"},
]
description = "CLI tool for YouTube playlist management"
readme = "README.md"
requires-python = ">=3.8"
keywords = ["youtube", "playlist", "cli", "downloader"]
license = {text = "MIT"}
classifiers = [
    "Development Status :: 4 - Beta",
    "Environment :: Console",
    "Intended Audience :: End Users/Desktop",
    "License :: OSI Approved :: MIT License",
    "Operating System :: OS Independent",
    "Programming Language :: Python :: 3",
    "Programming Language :: Python :: 3.8",
    "Programming Language :: Python :: 3.9",
    "Programming Language :: Python :: 3.10",
    "Programming Language :: Python :: 3.11",
    "Topic :: Multimedia :: Sound/Audio",
    "Topic :: Utilities",
]
dependencies = [
    "typer>=0.9.0",
    "yt-dlp>=2023.7.6",
    "google-api-python-client>=2.0.0",
    "google-auth-oauthlib>=1.0.0",
    "pyyaml>=6.0",
    "rich>=13.0.0",
]
dynamic = ["version"]

[project.optional-dependencies]
dev = [
    "pytest>=7.0.0",
    "pytest-cov>=4.0.0",
    "pytest-mock>=3.10.0",
    "black>=23.0.0",
    "flake8>=6.0.0",
    "mypy>=1.0.0",
    "pre-commit>=3.0.0",
    "tox>=4.0.0",
]
test = [
    "pytest>=7.0.0",
    "pytest-cov>=4.0.0",
    "pytest-mock>=3.10.0",
]

[project.urls]
Homepage = "https://github.com/cheroliv/playlist-downloader"
Documentation = "https://github.com/cheroliv/playlist-downloader#readme"
Repository = "https://github.com/cheroliv/playlist-downloader.git"
"Bug Tracker" = "https://github.com/cheroliv/playlist-downloader/issues"

[project.scripts]
playlist-downloader = "playlist_downloader.cli:main"

[tool.setuptools_scm]
write_to = "src/playlist_downloader/_version.py"

[tool.setuptools.packages.find]
where = ["src"]

[tool.pytest.ini_options]
testpaths = ["tests"]
python_files = ["test_*.py"]
python_classes = ["Test*"]
python_functions = ["test_*"]
addopts = [
    "--cov=src/playlist_downloader",
    "--cov-report=html",
    "--cov-report=term-missing",
    "--cov-fail-under=85",
]

[tool.black]
line-length = 88
target-version = ['py38']
include = '\.pyi?$'
extend-exclude = '''
/(
  \.eggs
  | \.git
  | \.hg
  | \.mypy_cache
  | \.tox
  | \.venv
  | _build
  | buck-out
  | build
  | dist
)/
'''

[tool.mypy]
python_version = "3.8"
warn_return_any = true
warn_unused_configs = true
disallow_untyped_defs = true
disallow_incomplete_defs = true
check_untyped_defs = true
disallow_untyped_decorators = true
no_implicit_optional = true
warn_redundant_casts = true
warn_unused_ignores = true
warn_no_return = true
warn_unreachable = true
strict_equality = true

[[tool.mypy.overrides]]
module = [
    "yt_dlp.*",
    "googleapiclient.*",
    "google_auth_oauthlib.*",
]
ignore_missing_imports = true

CI/CD Workflow - Tests and Quality

The main workflow (ci.yml) runs the tests on multiple versions of Python :

name: CI

on:
  push:
    branches: [ main, develop ]
  pull_request:
    branches: [ main ]

jobs:
  test:
    runs-on: ubuntu-latest
    strategy:
      matrix:
        python-version: ["3.8", "3.9", "3.10", "3.11"]

    steps:
    - uses: actions/checkout@v4
      with:
        fetch-depth: 0

    - name: Set up Python ${{ matrix.python-version }}
      uses: actions/setup-python@v4
      with:
        python-version: ${{ matrix.python-version }}

    - name: Cache dependencies
      uses: actions/cache@v3
      with:
        path: |
          ~/.cache/pip
          ~/.cache/pre-commit
        key: ${{ runner.os }}-pip-${{ hashFiles('**/requirements*.txt') }}
        restore-keys: |
          ${{ runner.os }}-pip-

    - name: Install dependencies
      run: |
        python -m pip install --upgrade pip
        pip install -e ".[dev]"

    - name: Lint with flake8
      run: |
        flake8 src tests --count --select=E9,F63,F7,F82 --show-source --statistics
        flake8 src tests --count --exit-zero --max-complexity=10 --max-line-length=88 --statistics

    - name: Check code formatting with Black
      run: black --check src tests

    - name: Type checking with mypy
      run: mypy src

    - name: Run tests with pytest
      run: |
        pytest tests/ -v --cov=src/playlist_downloader \
          --cov-report=xml --cov-report=term-missing

    - name: Upload coverage to Codecov
      uses: codecov/codecov-action@v3
      if: matrix.python-version == '3.11'
      with:
        file: ./coverage.xml
        flags: unittests
        name: codecov-umbrella

  security:
    runs-on: ubuntu-latest
    steps:
    - uses: actions/checkout@v4

    - name: Set up Python
      uses: actions/setup-python@v4
      with:
        python-version: "3.11"

    - name: Install dependencies
      run: |
        python -m pip install --upgrade pip
        pip install bandit[toml] safety

    - name: Run security checks with bandit
      run: bandit -r src/ -f json -o bandit-report.json

    - name: Check dependencies with safety
      run: safety check --json --output safety-report.json

    - name: Upload security reports
      uses: actions/upload-artifact@v3
      if: always()
      with:
        name: security-reports
        path: |
          bandit-report.json
          safety-report.json

  build:
    needs: [test, security]
    runs-on: ubuntu-latest
    steps:
    - uses: actions/checkout@v4
      with:
        fetch-depth: 0

    - name: Set up Python
      uses: actions/setup-python@v4
      with:
        python-version: "3.11"

    - name: Install build dependencies
      run: |
        python -m pip install --upgrade pip
        pip install build twine

    - name: Build package
      run: python -m build

    - name: Check package with twine
      run: twine check dist/*

    - name: Upload build artifacts
      uses: actions/upload-artifact@v3
      with:
        name: dist
        path: dist/

Release and Deployment Workflow

The release workflow (release.yml) manages the automatic publishing to PyPI :

name: Release

on:
  push:
    tags:
      - 'v*.*.*'
  workflow_dispatch:
    inputs:
      environment:
        description: 'Deployment environment'
        required: true
        default: 'test'
        type: choice
        options:
        - test
        - production

env:
  PYTHON_VERSION: "3.11"

jobs:
  release:
    runs-on: ubuntu-latest
    environment:
      name: ${{ github.event.inputs.environment || (startsWith(github.ref, 'refs/tags/') && 'production' || 'test') }}

    steps:
    - uses: actions/checkout@v4
      with:
        fetch-depth: 0

    - name: Set up Python
      uses: actions/setup-python@v4
      with:
        python-version: ${{ env.PYTHON_VERSION }}

    - name: Install dependencies
      run: |
        python -m pip install --upgrade pip
        pip install build twine

    - name: Build package
      run: python -m build

    - name: Check package
      run: twine check dist/*

    - name: Publish to Test PyPI
      if: github.event.inputs.environment == 'test' || (startsWith(github.ref, 'refs/tags/') && contains(github.ref, 'rc'))
      env:
        TWINE_USERNAME: __token__
        TWINE_PASSWORD: ${{ secrets.TEST_PYPI_API_TOKEN }}
      run: |
        twine upload --repository testpypi dist/*

    - name: Publish to PyPI
      if: github.event.inputs.environment == 'production' || (startsWith(github.ref, 'refs/tags/') && !contains(github.ref, 'rc'))
      env:
        TWINE_USERNAME: __token__
        TWINE_PASSWORD: ${{ secrets.PYPI_API_TOKEN }}
      run: |
        twine upload dist/*

    - name: Create GitHub Release
      if: startsWith(github.ref, 'refs/tags/')
      uses: actions/create-release@v1
      env:
        GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
      with:
        tag_name: ${{ github.ref }}
        release_name: Release ${{ github.ref }}
        draft: false
        prerelease: ${{ contains(github.ref, 'rc') }}

  post-release:
    needs: release
    runs-on: ubuntu-latest
    if: startsWith(github.ref, 'refs/tags/')

    steps:
    - uses: actions/checkout@v4

    - name: Set up Python
      uses: actions/setup-python@v4
      with:
        python-version: ${{ env.PYTHON_VERSION }}

    - name: Test installation from PyPI
      run: |
        sleep 60  # Attendre la propagation sur PyPI
        pip install playlist-downloader
        playlist-downloader --version

    - name: Update documentation
      run: |
        # Script pour mettre à jour la documentation
        echo "Documentation updated for version ${GITHUB_REF#refs/tags/}"

Architecture Diagrams

Sequence Diagram - Release Process

@startuml
!theme plain
title Release sequence to PyPI
actor Developer as dev
participant "GitHub" as gh
participant "GitHub Actions" as ga
participant "Test PyPI" as tpypi
participant "PyPI" as pypi
participant "Users" as users

dev -> gh : git push --tags v1.2.3
gh -> ga : trigger release workflow

ga -> ga : checkout code
ga -> ga : setup Python environment
ga -> ga : install dependencies
ga -> ga : run tests
ga -> ga : build package (wheel + sdist)
ga -> ga : check package with twine

alt Pre-release (rc tag)
    ga -> tpypi : upload to Test PyPI
    tpypi -> ga : confirm upload
else Stable release
    ga -> pypi : upload to PyPI
    pypi -> ga : confirm upload
end

ga -> gh : create GitHub release
ga -> ga : test installation from PyPI

users -> pypi : pip install playlist-downloader
pypi -> users : download package

@enduml

State Diagram - Package Life Cycle

@startuml
!theme plain
title States of the Python Package
[*] --> Development
Development --> Testing : commit/PR
Testing --> Development : tests fail
Testing --> Built : tests pass
Built --> TestPyPI : pre-release tag
Built --> PyPI : stable tag
TestPyPI --> PyPI : validation OK
PyPI --> Published
Published --> [*]

state Development {
  [*] --> Coding
  Coding --> LocalTesting
  LocalTesting --> Coding : fix issues
  LocalTesting --> ReadyForCI : all tests pass
}

state Testing {
  [*] --> CITests
  CITests --> SecurityScan
  SecurityScan --> QualityCheck
  QualityCheck --> BuildValidation
}

@enduml

Deployment Diagram - CI/CD Infrastructure

@startuml
!theme plain
title Deployment Infrastructure
node "GitHub" {
  component "Repository" as repo
  component "Actions Runner" as runner
  component "Secrets Store" as secrets
}

node "PyPI Infrastructure" {
  component "PyPI" as pypi
  component "Test PyPI" as testpypi
  database "Package Index" as index
}

node "Developer Machine" {
  component "Git Client" as git
  component "Python Environment" as python
  component "IDE" as ide
}

node "User Environment" {
  component "pip" as pip_client
  component "Python Runtime" as py_runtime
}

git --> repo : push code/tags
repo --> runner : trigger workflows
runner --> secrets : read API tokens
runner --> testpypi : upload pre-release
runner --> pypi : upload release
pypi --> index : store package
pip_client --> pypi : download package
py_runtime <-- pip_client : install package

@enduml

Objects and Models of the Pipeline

Class Diagram - CI/CD Models

@startuml
!theme plain
title Models of the CI/CD pipeline
class PipelineConfig {
  +python_versions: List[str]
  +test_environments: List[str]
  +security_checks: bool
  +coverage_threshold: float
  +validate()
}

class BuildArtifact {
  +name: str
  +version: str
  +wheel_path: str
  +sdist_path: str
  +checksums: Dict[str, str]
  +validate_integrity()
}

class TestResult {
  +test_suite: str
  +python_version: str
  +passed: int
  +failed: int
  +coverage: float
  +duration: float
  +is_success(): bool
}

class SecurityReport {
  +bandit_issues: List[Issue]
  +safety_vulnerabilities: List[Vulnerability]
  +severity_level: str
  +is_secure(): bool
}

class DeploymentTarget {
  +name: str
  +url: str
  +api_token: str
  +environment: str
  +deploy(artifact: BuildArtifact)
}

class ReleaseManager {
  +version: str
  +changelog: str
  +artifacts: List[BuildArtifact]
  +test_results: List[TestResult]
  +security_report: SecurityReport
  +deploy_to_test()
  +deploy_to_production()
  +create_github_release()
}

PipelineConfig ||--o{ TestResult
ReleaseManager *-- BuildArtifact
ReleaseManager *-- SecurityReport
ReleaseManager o-- DeploymentTarget
DeploymentTarget ..> BuildArtifact : uses

@enduml

Secrets Configuration

For the pipeline to work, you must configure the following secrets in GitHub:

GitHub Actions secrets

# Dans Settings > Secrets and variables > Actions

# Token PyPI pour la production
PYPI_API_TOKEN=pypi-...

# Token Test PyPI pour les pré-releases
TEST_PYPI_API_TOKEN=pypi-...

# Token GitHub pour créer les releases
GITHUB_TOKEN=(automatiquement fourni)

# Token Codecov (optionnel)
CODECOV_TOKEN=...

Generation of PyPI Tokens

# 1. Créer un compte sur PyPI et Test PyPI
# 2. Aller dans Account Settings > API tokens
# 3. Créer un token avec scope "Entire account" ou spécifique au projet
# 4. Format du token : pypi-AgEIcHlwaS5vcmc...

Local Development Scripts

To facilitate development, create utility scripts:

Makefile

.PHONY: install test lint format security build clean release-test release-prod

install:
	pip install -e ".[dev]"

test:
	pytest tests/ -v --cov=src/playlist_downloader

lint:
	flake8 src tests
	mypy src

format:
	black src tests

security:
	bandit -r src/
	safety check

build:
	python -m build
	twine check dist/*

clean:
	rm -rf build/ dist/ *.egg-info/
	find . -type d -name __pycache__ -delete
	find . -name "*.pyc" -delete

release-test: clean build
	twine upload --repository testpypi dist/*

release-prod: clean build
	twine upload dist/*

pre-commit: format lint test security
	@echo "✅ Prêt pour commit"

Version script

#!/usr/bin/env python3
"""Script pour gérer les versions du projet."""

import sys
import subprocess
from pathlib import Path

def get_current_version():
    """Récupère la version actuelle depuis git."""
    try:
        result = subprocess.run(
            ["git", "describe", "--tags", "--abbrev=0"],
            capture_output=True,
            text=True,
            check=True
        )
        return result.stdout.strip()
    except subprocess.CalledProcessError:
        return "0.0.0"

def create_version_tag(version, message=None):
    """Crée un tag de version."""
    if not version.startswith('v'):
        version = f'v{version}'

    tag_message = message or f"Release {version}"

    subprocess.run(["git", "tag", "-a", version, "-m", tag_message], check=True)
    print(f"✅ Tag {version} créé")

    # Push le tag
    subprocess.run(["git", "push", "origin", version], check=True)
    print(f"✅ Tag {version} poussé vers origin")

if __name__ == "__main__":
    if len(sys.argv) < 2:
        current = get_current_version()
        print(f"Version actuelle: {current}")
        print("Usage: python version.py <new_version> [message]")
        sys.exit(1)

    new_version = sys.argv[1]
    message = sys.argv[2] if len(sys.argv) > 2 else None

    create_version_tag(new_version, message)

Best Practices and Recommendations

Semantic versioning

Use semantic versioning (SemVer) :

  • MAJOR.MINOR.PATCH (ex: 1.2.3)

  • `MAJOR`incompatible changes

  • MINOR: new compatible features

  • `PATCH`compatible bug fixes

Branching Strategy

main          ──●──●──●──●──●────●── (releases stables)
               /       /          /
develop    ──●──●──●──●──●──●──●──●── (développement)
            /     /        /
feature/xxx  ●──●──●──●──●──/ (fonctionnalités)

Tests and Coverage

  • Minimum code coverage: 85%

  • Unit tests for the business logic

  • Integration tests for the adapters

  • End-to-end tests for the CLI

Security

  • Automatic dependency scanning (Safety)

  • Static code analysis (Bandit)

  • Secrets are never in the code

  • Using specific PyPI tokens

Monitoring and Observability

Pipeline metrics

# .github/workflows/metrics.yml
name: Pipeline Metrics

on:
  workflow_run:
    workflows: ["CI", "Release"]
    types: [completed]

jobs:
  metrics:
    runs-on: ubuntu-latest
    steps:
    - name: Collect metrics
      run: |
        echo "Pipeline: ${{ github.event.workflow_run.name }}"
        echo "Status: ${{ github.event.workflow_run.conclusion }}"
        echo "Duration: ${{ github.event.workflow_run.updated_at - github.event.workflow_run.created_at }}"
        # Envoyer vers système de monitoring

Notifications

# Ajout dans les workflows pour notifications
- name: Notify on failure
  if: failure()
  uses: 8398a7/action-slack@v3
  with:
    status: failure
    text: "❌ Pipeline failed for ${{ github.repository }}"
  env:
    SLACK_WEBHOOK_URL: ${{ secrets.SLACK_WEBHOOK }}

Conclusion

This complete CI/CD pipeline for Python provides:

  • Full automationfrom code validation to publication

  • Security: automatic scans and secure secret management

  • Qualitymulti-version tests, linting and code coverage

  • Reliability: progressive deployment via Test PyPI

  • Traceability: artifacts, reports and releases GitHub

Adopting these practices ensures a robust and professional delivery process for your Python CLI applications, making it easier to maintain and evolve your projects in the long term.

Related articles