What is uv
I’ve recently started using uv to manage Python projects and packages: many projects at work and the one active package I have on pypi.org (habanero).
I don’t really know enough about all the various Python tools similar to uv to give an informed opinion. Rather, this is purely reflections on using uv.
uv’s tagline is:
An extremely fast Python package and project manager, written in Rust.
uv docs and source code.
Reflections
Fast
It’s super fast. I don’t remember my experience with plain ol pip being anywhere near this fast, though surely that could be user (me) incompetence.
Self python
The ability for uv to install Python itself is very nice. This means I don’t have to manage downloading Python myself, and worry about adding things to the path, etc.
User land
With uv I do not run into user vs. not user install issues I had with pip often. This was probably just me not using python/pip the right way? Anyway, uv doesn’t support the --user flag it turns out.
uv run
It’s very nice that uv run enters and exits the virtual environment for you. This wasn’t super obvious that this was even happening when I first starting using the command, but of course makes sense now. You can still activate/deactivate the virtual env yourself though.
Inline script metadata
uv supports a concept of inline script metadata - which comes from PEP 723. Simon Willison also wrote about this recently.
It looks like (example from their docs)
# /// script
# dependencies = [
# "requests<3",
# "rich",
# ]
# ///
import requests
from rich.pretty import pprint
resp = requests.get("https://peps.python.org/api/peps.json")
data = resp.json()
pprint([(k, v["title"]) for k, v in data.items()][:10])
I have not used this feature but I think it could be very helpful. There’s lots of Python scripts that really don’t need a project structure but would benefit from properly managing dependencies. Now with this feature there’s no need to have the heavier project structure when you have all your code in a single script.
REPL tools
I don’t know if this is the “right” way to do things, but there’s some dependencies my project or package do not depend on but that I want to use for development. I like to use ipython repl instead of the default python one - although the one that comes with Python is getting very good. So I install ipython and rich with uv pip install ipython rich. This invocation does not add them as dependencies but still (I think?) manages these dependencies within the project/package structure.
VC backed?
I’m a little nervous about uv being made by a VC backed company. If I come to depend on uv and then they get bought and things change, or they just fold soon after, that’s a bummer. But, it is open source so I imagine projects as big and important as uv and ruff may be adopted by the community. And I absolutely want people making open source software to be able to make a living, so we kinda need to keep our options open for how to do that.
Links
Other blog posts covering uv: