Python¶

Python development setup and package management using modern tools.

Overview¶

This setup uses two modern package managers: - UV: Fast, reliable Python package manager and project tool - Pixi: Cross-platform package manager compatible with conda

Installing Indigo (Python Binary Distribution)¶

For a clean, optimized Python environment, use Indigo - a standalone Python distribution.

Download and Install¶

Download the latest Indigo release

curl -L https://github.com/indygreg/python-build-standalone/releases/download/20240713/cpython-3.13.0b4+20240713-x86_64-unknown-linux-gnu.tar.zst -o python.tar.zst

Extract to your home directory

mkdir -p ~/.python
tar -I zstd -xf python.tar.zst -C ~/.python

Add to PATH

Add this to your shell configuration (.bashrc, .zshrc):

export PATH="$HOME/.python/python/install/bin:$PATH"

Verify Installation

1	`python --version`

UV Package Manager¶

UV is a fast, reliable package manager for Python projects.

Installation¶

curl -LsSf https://astral.sh/uv/install.sh | sh

Creating a New Project¶

uv init my-project
cd my-project

Adding Dependencies¶

uv add requests
uv add --dev pytest

Running Python Scripts¶

1	`uv run script.py`

Installing from pyproject.toml¶

uv sync

Creating a Virtual Environment¶

uv venv
. .venv/bin/activate  # On Windows: .venv\Scripts\activate

Pixi Package Manager¶

Pixi is a cross-platform package manager compatible with conda environments, great for data science and scientific computing.

Installation¶

curl -fsSL https://pixi.sh/install.sh | bash

Creating a New Project¶

pixi init my-data-project
cd my-data-project

Adding Dependencies¶

pixi add numpy pandas
pixi add --pypi requests

Running Commands¶

pixi run python script.py

Creating Environments¶

pixi create -e analysis
pixi run -e analysis python analysis.py

Conda Compatibility¶

Pixi automatically handles conda channels and environments:

pixi add -c conda-forge numpy

Choosing Between UV and Pixi¶

Use Case	Tool
Standard Python projects	UV
Data science / Jupyter	Pixi
Machine learning	Pixi
Web development	UV
CLI tools	UV
Scientific computing	Pixi
Mixed dependencies (PyPI + conda)	Pixi

Virtual Environments¶

With UV¶

# Create virtual environment
uv venv

# Activate
. .venv/bin/activate

# Deactivate
deactivate

With Pixi¶

# Create environment
pixi create -e myenv

# Use environment
pixi run -e myenv python script.py

Best Practices¶

Always use virtual environments to isolate project dependencies
Use UV for most Python projects - it's faster and simpler
Use Pixi for data science - better conda ecosystem support
Commit lock files (uv.lock or pixi.lock) to version control
Document dependencies in pyproject.toml or pixi.toml