-
-
Notifications
You must be signed in to change notification settings - Fork 467
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
support dependency groups #421
Comments
It also would be great to be able to build wheels by specifying dependency groups. I want to build a CPU version of my package that uses the CPU version of pytorch, and another that uses the GPU version and also installs a few other packages. poetry doesn't support this even though it has dependency groups. |
Came here to say it would be nice to see this feature. Although I'm a bit confused as to what do we mean by dependency groups.
If we're referring to extras, I think rye already supports them. To add
To add
What's confusing is that if you need to do
|
This actually doesn't work for me, must have something to do with a limitation in virtual packages (#639) |
It would be pretty cool if Rye produced a separate lockfile for each dependency group. I.e. if we have a dependency group Then in my RUN uv pip install -r requirements-foobar.lock |
Absolutely. Right now I resort to hacky solutions managing multiple pyproject files (or lock files) for each env (e.g. env with jax and another with torch -- these install different cuda dependencies so they don't play well with each other) |
FYI. astral-sh/uv#8090 I think this can be closed as people/new features are moving to |
I think this is a very convenient feature that
pdm
andpoetry
has. When I try to make a package that may use CUDA, I need to have a dependency that depends on hardware configurations, and a package manager has no means to know that thus I believe the best way is to let users decide which group of dependencies they should install. This is like the features at the package level.Currently, there is no way to specify both CPU and GPU dependencies with rye within one
pyproject.toml
, cargo also has something similar that allows one to specify dependencies based on the current crates' features. And I don't see a way of passing "extra"s in python'spyproject.toml
(e.g pass the extra[cuda12]
across packages) I hope this could be something rye can fixThe text was updated successfully, but these errors were encountered: