A couple of months ago I made the switch from poetry to uv. I’m writing the post just to say i love it 💚, and you should use it too! Here’s how it impacts my development expericence and how I use it to quickstart new Python projects.

TLDR; uv simplifies my whole development experience. I use it with my own cookiecutter template – available here:

GitHub

What is uv?

If you’re reading this, you probably already know uv. But just in case, here’s a quick summary:

uv is a package management tool that allows you manage your Python installations. It’s main selling point is its speed since uv is written in Rust. And fast it is! Nonetheless, it has a lot more features to offer.

Why I switched to uv

I was using poetry, it’s a great tool, don’t get me wrong. Then I discovered uv, was curious and simply gave it a try. Soon I ‘exclusively’ set-up new projects with uv. Later, I even started to migrate my existing projects to uv. Since I was already used to the pyproject.toml file for project configuration, the migration was easy. Along the way I discovered a couple of features that impact my workflow positively:

1. Set-up a machine for Python development

There is no need to install Python manually anymore. On a new machine, I simply run the installation command for uv and I’m good to go.

With another

uv python install

command I can install the latest Python version. That’s it for the set-up process.

In fact, I recommend the same approach to my colleagues. Setting up an existing project on a colleague’s machine has never been easier for me. It’s a simple three step process:

1. Install uv
2. Clone the repository
3. Run uv sync

Even if my colleague does not have the appropriate Python version installed, uv will take care of that. It will download the correct version and set up a virtual environment for the project.

This way setting up a project is a matter of a couple of minutes.

2. Switching between Python versions

Sometimes I need to switch between Python versions. Previously, I had the manage the Python version manually which lead to a couple of issues. Now, I don’t hesitate and I feel encouraged to test my code with different Python versions.

3. Installing tools

I like some tools like ruff and cookiecutter to be installed system-wide. With uv, I can install them with a single command:

uv tool install ruff

4. Documentation

Sometimes it’s a point overlooked, but I think it’s important to mention: uv has great documentation. It’s easy to find the information you need. Especially, the Integration section was practically a quickstart guide to using uv with Docker and pre-commit. So I was able to easily integrate uv into my projects.

5. Speed

… and of course, uv is fast. Compared to poetry, resolving dependencies is way faster, which reduces friction in my workflow. For example, I am more likely to try out a new package or quickly set up a new environment just to test something out.

How I use it day-to-day

For new projects, I created my own template to quickly set up a new Python project. It is available on GitHub.

GitHub

The template is centered around uv and includes a suitable project structure and tools for common tasks like testing, linting, and formatting.

If you’d like to use it, follow these three simple steps:

1. Install uv (Installation instructions here)
2. Install pre-commit and cookiecutter as tools:
   uv tool install pre-commit cookiecutter
3. Create a new project with the template:

cookiecutter https://github.com/JakobKlotz/python-template.git

Simple as that! 🚀

So there you have it – uv has transformed my Python workflow with its speed, simplicity, and seamless version management. Give it a try, and I’m confident it will enhance your development experience just as it has mine!