kubeflow / pipelines

Machine Learning Pipelines for Kubeflow
https://www.kubeflow.org/docs/components/pipelines/
Apache License 2.0
3.5k stars 1.58k forks source link

[sdk] Relax `python_requires` in `setup.py` #10738

Open AndersBennedsgaard opened 2 months ago

AndersBennedsgaard commented 2 months ago

Environment

Steps to reproduce

I tried to install kfp using Poetry, where I've set the allowed Python versions to python = "^3.11.0". However, this results in an version solving failed error since 3.13.0 is allowed in that range, but isn't allowed by kfp:

https://github.com/kubeflow/pipelines/blob/c967d9f7df0bec5827cdf45ea02d3463d8b17aff/sdk/python/setup.py#L102

As new minor versions of Python should be backwards compatible, there is no real reason for setting an upper limit to the Python version to <3.13.0.

Expected result

poetry add kfp doesn't return an error, since python_requires='>=3.7.0' is used

As a workaround I can set kfp = { version = "^2.7.0", python = ">=3.7,<3.13.0" } to force Poetry to figure out the highest allowed Python version, but we shouldn't have to do this.

Materials and Reference


Impacted by this bug? Give it a 👍.

github-actions[bot] commented 1 week ago

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

AndersBennedsgaard commented 1 week ago

Still relevant