You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I searched open requests and couldn't find a duplicate
What is the idea?
Apologies if I've missed this feature in the readme or prior discussion in the issues.
I'm helping someone setup conda-lock with a pyproject.toml setup that includes:
[project.optional-dependencies]
cuda = [
"cupy-cuda11x ; platform_system != 'Darwin'",
]
naturally, specific cuda builds of of cupy aren't declared on conda as they are on pypi, but grayskull's lookup doesn't (currently) deal with all of those maps.
I know that I can provide my own --pypi_to_conda_lookup_file, but that appears to A) require a URI I can easily pass to requests.get B) presumably requires me to duplicate stuff that I need from grayskull's map?
looking through the source code, there doesn't appear to be any way to append to the lookup.
Why is this needed?
to assist in pypi-to-conda name mapping with minimal duplication
[tool.conda-lock.pypi-to-conda-name]
cupy-cuda11x = "cupy"# or multiple packages, as would be useful in the case of cupycupy-cuda11x = ["cupy", "cuda-version=11.8"]
Additional Context
No response
The text was updated successfully, but these errors were encountered:
Checklist
What is the idea?
Apologies if I've missed this feature in the readme or prior discussion in the issues.
I'm helping someone setup conda-lock with a pyproject.toml setup that includes:
naturally, specific cuda builds of of cupy aren't declared on conda as they are on pypi, but grayskull's lookup doesn't (currently) deal with all of those maps.
I know that I can provide my own
--pypi_to_conda_lookup_file
, but that appears to A) require a URI I can easily pass to requests.get B) presumably requires me to duplicate stuff that I need from grayskull's map?looking through the source code, there doesn't appear to be any way to append to the lookup.
Why is this needed?
to assist in pypi-to-conda name mapping with minimal duplication
What should happen?
I'm thinking something like:
or simpler if no additional info is needed
Additional Context
No response
The text was updated successfully, but these errors were encountered: