Add constraints to meta.yaml

conda-forge / pandas-feedstock

A conda-smithy repository for pandas.

BSD 3-Clause "New" or "Revised" License

8 stars 35 forks source link

Add constraints to meta.yaml #212

Open jefsayshi opened 1 month ago

jefsayshi commented 1 month ago

Checklist

[x] Used a personal fork of the feedstock to propose changes
[ ] Bumped the build number (if the version is unchanged)
[ ] Reset the build number to 0 (if the version changed)
[x] Re-rendered with the latest conda-smithy (Use the phrase code>@<space/conda-forge-admin, please rerender in a comment in this PR for automated rerendering)
[ ] Ensured the license file is being packaged.

Resolves #187

Summary of Changes:

align with constraints on conda main channel
https://pandas.pydata.org/docs/dev/whatsnew/v2.2.0.html#increased-minimum-versions-for-dependencies

conda-forge-webservices[bot] commented 1 month ago

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe/meta.yaml) and found it was in an excellent condition.

jefsayshi commented 1 month ago

@conda-forge-admin, please rerender

jefsayshi commented 3 weeks ago

@h-vetinari or @lithomas1 is there anything that I need to do to get this merged? I would prefer that it gets merged before another pandas version is released so that the run constraints will be added to it.

h-vetinari commented 3 weeks ago

I'm not up to date with the extra-installs that pandas offers. The constraints are probably an OK fix, though it reduces the amount of legal environments that pandas can be installed in.

Better would be to create equivalent outputs here where e.g. pip install pandas[performance] is mapped to conda install pandas-performance. There are many examples of feedstocks that do this, check out for example ray or gymnasium.

But ultimately you'll have to figure this out with the maintainers of this feedstock, which doesn't include me; I'm just helping out here and there.

jamesmyatt commented 3 weeks ago

jefsayshi commented 3 weeks ago

I agree that it will reduce the number of solves, but the ones reduced are only legal until you try to use a pandas function that relies on one of them. My example in issue #187 is sqlalchemy 1.4 and pandas 2.2

This PR will at least align conda-forge with the conda main channel.

While subpackages could end up being the long term solution, there does not seem to be consensus around it yet.

lithomas1 commented 3 weeks ago

I'm taking a break from pandas maintenance this summer.

I'll cycle back to this in a couple weeks, please ping if I don't. (This looks correct at a first glance, though.) Thanks.

lithomas1 commented 1 week ago

Looking again, I think this just needs a bump of the build number.

(CI is red at the moment, so this wouldn't be mergeable until the next release anyways so maybe the bump is unnecessary and we just wait for the next release)

jefsayshi commented 1 week ago

I am fine with waiting for the next release, which is why I didn't bump the build number. However, I will defer to your preference, so just let me know what you want me to do.