Python Packaging

Modern Python packaging with pyproject.toml, uv, and best practices for distribution.

Quick Start

# Create new project with uv
uv init my-package
cd my-package

# Add dependencies
uv add requests click

# Build package
uv build

# Publish to PyPI
uv publish

When to Use

• Creating distributable Python libraries
• Building CLI tools
• Publishing to PyPI
• Setting up development environments
• Managing project dependencies

Core Decisions

1. Layout Choice

# Source layout (recommended)
src/my_package/
    __init__.py
    module.py

# Flat layout (simple)
my_package/
    __init__.py
    module.py

Source layout benefits:

• Clear separation of source and tests
• Prevents accidental imports of uninstalled code
• Better for packages with complex structure

2. Project Structure

Minimal Project:

my-project/
├── pyproject.toml
├── README.md
├── src/
│   └── my_package/
│       └── __init__.py
└── tests/
    └── test_init.py

Complete Project:

my-project/
├── pyproject.toml
├── README.md
├── LICENSE
├── .gitignore
├── src/
│   └── my_package/
│       ├── __init__.py
│       ├── cli.py
│       ├── core.py
│       └── utils.py
├── tests/
│   ├── conftest.py
│   └── test_core.py
└── docs/
    └── index.md

Detailed Topics

See modules for detailed information:

• uv Workflow - Complete uv commands and troubleshooting
• pyproject.toml Patterns - Configuration examples for different package types
• Entry Points - Console scripts, GUI scripts, and plugins
• CI/CD Integration - GitHub Actions and automated publishing

Best Practices

1. Use source layout for anything beyond simple packages
2. Pin direct dependencies with minimum versions
3. Use optional dependency groups for dev/docs/test
4. Include py.typed for type hint support
5. Add comprehensive README with usage examples
6. Use semantic versioning (MAJOR.MINOR.PATCH)
7. Test on multiple Python versions before publishing

Exit Criteria

• Modern pyproject.toml configuration
• Clear dependency specification
• Proper version management
• Tests included and passing
• Build process reproducible
• Publishing pipeline automated