autopep723 0.1.0

Auto-generate PEP 723 metadata for Python scripts by analyzing imports

autopep723

A CLI tool that automatically generates PEP 723 metadata for Python scripts by analyzing their imports and dependencies.

What it does

autopep723 parses Python scripts to detect third-party package imports and generates the appropriate PEP 723 inline script metadata. This metadata allows tools like uv run to automatically install and manage dependencies when executing scripts.

Installation

You can run the tool directly using uvx:

uvx autopep723

To install it permanently:

uv tool install autopep723

Quick Start

# Run script with automatically detected dependencies
autopep723 script.py

# Print PEP 723 metadata to stdout  
autopep723 check script.py

# Update the script file with metadata
autopep723 upgrade script.py

Shebang Integration

You can use autopep723 directly as a shebang:

#!/usr/bin/env uvx autopep723
import requests
import numpy as np

# Your script here...

This allows scripts to be executable without explicitly declaring dependencies. The tool detects imports and runs the script using uv run with the required packages installed on-the-fly in an ephemeral environment.

Features

• 🔍 Automatic dependency detection via AST analysis
• 📦 Import name mapping (e.g., import PIL → pillow)
• 🚀 Multiple output modes (print, update, or run)
• ✅ PEP 723 compliant metadata generation
• 🛡️ Built-in module filtering excludes standard library
• 🔧 Graceful error handling for syntax errors

Documentation

For detailed usage, examples, and API reference, see the full documentation.

Related

• PEP 723 - Inline script metadata specification
• uv tools guide - Using uv for tool management
• uv issue #6283 - Autodetect dependencies mode proposal

Contributing

Pull requests are welcome! 🎉

• One change per PR - Keep it focused
• Tests required - Maintain coverage standards
• Follow conventions - Use pytest and pytest-mock
• Be thoughtful - This is a minimal wrapper, new dependencies need strong justification

AI assistance is encouraged (this project is ~100% AI-generated), but contributions must meet quality standards.

We follow a zero-tolerance policy for harassment of any kind. Be respectful and inclusive.

License

MIT - see LICENSE file for details.