I’ve been using some new Python tools lately, uv instead of venv+pip, and marimo instead of jupyterlab/ipynotebooks. I love jupyterlab, it’s how I learned to program 15 years ago, via jupyter notebooks, and ipython, but as the Python versions have piled up over the years, and all the different solutions for versions and environments have cost me hours and hours over the years, it’s so nice to find a tool like uv that Just Works, and a notebook environment like marimo, that works with uv and also Just Works. Marimo has mroe features than you can shake a stick at, but it’s dayenu feature is that you can create an environment and run a notebook from that environment. I never have figured out how to do that with jupyter–uv and marimo Just Work.

uv

Installation: brew install uv. That’s it.

Create project with: uv init project-name.

Run a project’s entrypoint: uv run main.py.

Add a module: uv add pandas, or remove: uv remove module.

In this case, all of your top level modules are stored in a pyproject.toml.

If you need a requirements.txt file, then use the command uv pip compile pyproject.toml -o requirements.txt. This will lock the dependencies specified in your pyproject.toml and write them to the requirements.txt file.

If you need to pull from a specific branch from a GitHub, then you can do that like this,

uv add git+https://github.com/<username>/<project> --branch main

marimo

You can use uv to install things globally using their “tool” subcommand: uv tool install tool-name. I’ll install marimo with it’s recommended extras. uv tool install "marimo[recommended]"