poetry-packaging_skill

Overview

This skill teaches modern Python package management and publishing with Poetry. It covers creating projects, managing reproducible dependencies, building distributions, and publishing to TestPyPI/PyPI. You’ll learn project layout, semantic versioning, and workflows for reliable releases.

How this skill works

The skill walks through Poetry commands and pyproject.toml configuration to define metadata, dependency groups, and build settings. It demonstrates dependency resolution, lock file maintenance, virtual environment usage, and building sdist/wheel artifacts. You’ll practice publishing flow, tagging releases, and integrating packaging into CI/CD pipelines.

When to use it

• Starting a new Python library or CLI that you plan to publish
• Standardizing dependency management and reproducible builds across projects
• Preparing packages for PyPI or a private repository
• Implementing CI/CD pipelines for automated releases
• Managing monorepos or projects with multiple dependency groups

Best practices

• Keep pyproject.toml authoritative: declare metadata, dependencies, and scripts there
• Use dependency groups for dev, test, and docs to avoid bloating installs
• Pin or use appropriate version constraints (caret/tilde) and commit poetry.lock for reproducibility
• Test publishing to TestPyPI before releasing to PyPI
• Automate builds and publishes in CI using encrypted PyPI tokens and git tags

Example use cases

• Create a CLI package with Click/Typer, expose entry points, and publish to PyPI
• Develop a reusable library with type hints, tests, and documentation hosted on Read the Docs
• Set up a monorepo with multiple packages, shared dependency groups, and a CI release pipeline
• Build a plugin-based package using entry points and distribute example plugins
• Configure platform-specific dependencies and optional extras for database drivers