Skip to content

📦 Package the GitLab API Wrapper Project Using Astral's uv

This issue tracks the work required to properly package our in-house GitLab API wrapper as a Python project using Astral's uv for modern dependency management and reproducible environments.

🎯 Goals:

  • Package the project into a proper Python module
  • Use uv to handle virtual environment creation, dependency resolution, and lockfile management
  • Enable easy local development, distribution, and CI/CD integration

✅ Tasks:

1. Project Structure

Ensure the codebase follows a standard layout:

gitlab_api_wrapper/
├── src/
│   └── gitlab_wrapper/   # Main package
├── tests/                # Unit tests
├── pyproject.toml        # Build config
└── README.md

2. Add pyproject.toml

Configure the project for packaging with PEP 621. Include metadata, dependencies, and build system.

[project]
name = "gitlab-api-wrapper"
version = "0.1.0"
description = "In-house GitLab API wrapper"
authors = [{name = "Your Org", email = "[email protected]"}]
dependencies = [
    "python-gitlab",
    # other deps
]

[build-system]
requires = ["setuptools", "wheel"]
build-backend = "setuptools.build_meta"

3. Initialize uv

  • Run uv init to generate the uv virtual environment and lockfile
  • Ensure .gitignore includes .venv/ and uv.lock 
uv init
uv pip install -e .
uv pip install pytest  # or other dev/test tools
uv venv .venv

4. Testing and Linting

  • Add basic pytest setup
  • DO formatting and linting with ruff

5. CI/CD Compatibility

  • Update GitLab CI jobs to use uv for setting up environment (if applicable)
  • Replace existing pip or venv steps in .gitlab-ci.yml with uv

📝 Notes:

  • uv is significantly faster and more deterministic than pip + venv, with lockfile support and zero external dependencies
  • This will make the project easier to install and maintain for both internal developers and automation

📌 References:

Edited by Ranjith Raj