Fast Python Development with uv

Details

Format

This tutorial is highly interactive and hands-on:

• Live Demonstrations (approximately 60%): I develop examples step-by-step, starting from blank notebooks or terminal sessions. Participants are encouraged to type along. The supplied PDF contains all code to help participants get back on track if needed.
• Exercises (approximately 35%): Three exercise sessions where participants apply concepts to their own projects or work through structured examples. Solutions are reviewed together.
• Discussion and Q\&A (approximately 5%): Questions are encouraged at any time. I explicitly ask for feedback on pacing and clarity throughout.

Teaching approach:

• No pre-filled notebooks—everything is built live to show the thought process
• Typing speed is deliberately moderate to allow participants to follow
• More complex examples are also provided as Python source code files
• Comprehensive PDF handout with all content for reference
• Solutions PDF distributed after the tutorial

Participants will be active throughout the entire tutorial, working with uv on their own machines to build real-world skills.

Outline

Introduction (5 min)

• Instructor introduction
• Participant background survey (raise hands to answer questions about Python and dependency management experience)

Problem Analysis (10 min)

Explain what dependency management is needed for.

Two basic tasks:

• Installing packages
• Building packages

Questions to set the stage:

• Why is dependency management complex?
• What approaches exist?
• Why is speed important?

Overview of the Ecosystem (15 min)

Look at available tools and explain key terms.

• Traditional pip-based approaches
  • pip and pip-tools
  • virtualenv and venv
  • pyenv for Python version management
  • poetry
  • pdm
  • hatch/hatchling
  • setuptools
• conda-based approaches (brief overview)
  • conda/mamba
  • Pixi (hybrid approach with conda-forge + PyPI)
• Modern unified approach
  • uv: one tool to replace them all

The uv Approach (30 min)

• Command-line interface (CLI) and pip compatibility
• pyproject.toml configuration file
• uv.lock for reproducibility
• Virtual environments (automatic management)
• Python version management with uv python
• Global tool installation with uv tool and uvx
• Single-file scripts with inline dependencies
• Exercise (10 min): Getting started with uv

Example - A Small Python Data Project (40 min)

• Define an example: a small program to analyze data for internal use
• Start with pandas
• Add more dependencies progressively:
  • openpyxl for Excel file support
  • matplotlib for visualization
  • pyarrow for Parquet files
• Use pyproject.toml and uv.lock to manage dependencies
• Compare installation speed with pip
• Address potential problems and conflicts
• Exercise (15 min): Apply uv to your own program

Example - Building a Python Package (40 min)

• Turn the example into a distributable package
• Use pyproject.toml for package metadata
• Use editable installs for development
• Configure multiple environments for:
  • Testing with pytest
  • Different Python versions
  • Documentation building
• Workspaces for multi-package projects
• Building and publishing packages
• Exercise (15 min): Package your own project

Advanced Features and Best Practices (30 min)

• Using uvx for running tools without installation
• Single-file scripts with inline dependency metadata
• Migrating from pip, poetry, or other tools
• Working with requirements.txt files
• Special cases:
  • Handling platform-specific dependencies
  • Integrating with CI/CD pipelines
• Performance tips and troubleshooting
• Brief comparison: uv vs. Pixi vs. other tools

Wrap-Up and Q&A (10 min)

• When to use uv (and when to consider alternatives like Pixi)
• Migration strategies from other tools
• Resources for continued learning
• Open questions and discussion

Internet Requirements (Prep for Tutorial)

Participants will receive detailed setup instructions a few days before the tutorial, including:

• uv installation instructions for Windows, macOS, and Linux
• Verification steps to ensure uv is working correctly
• PDF handout with tutorial content
• Code examples and exercises
• Pre-downloading instructions for common packages (optional but recommended)

Critical: Participants MUST have uv installed before the tutorial.

Internet Requirements (During Tutorial)

Internet connection required: This tutorial requires internet connectivity during the session because:

• Package downloads are a core part of the tutorial (installing packages with uv)
• Participants will download packages from PyPI
• Live demonstrations show real-world package installation scenarios
• Speed comparisons require actual downloads

Bandwidth considerations: To minimize bandwidth usage:

• Because uv caches packages, participants can reduce bandwidth usage during the tutorial by installing packages a day before the tutorial
• I will provide instructions how to do this
• uv's efficient caching means repeated installs use minimal bandwidth

Fallback plan: If internet becomes unavailable:

• I can still show how my pre-installed environments work
• The tutorial can pivot to focus on uv configuration and workflow management