try removing py3.9 limitation

[rabernat]

Checklist

• 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)
• Re-rendered with the latest conda-smithy (Use the phrase @conda-forge-admin, please rerender in a comment in this PR for automated rerendering)
• Ensured the license file is being packaged.

Thanks so much for maintaining this recipe! 🙏

I see that this recipe doesn't support python 3.9

apache-beam-feedstock/recipe/meta.yaml

Lines 14 to 15 in 5f83ff5

	# py39 versions of python-avro and dill are too recent
	skip: true # [win or py>=39]

However, Beam support for python 3.9 overall seems solid now (see https://issues.apache.org/jira/browse/BEAM-12000?focusedCommentId=17541173&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-17541173). Therefore, I am naively trying to remove the python 3.9 limitation to see if we can get this build working for 3.9.

Thanks for your patience. I'm new to Beam.

[conda-forge-linter]

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) and found it was in an excellent condition.

[rabernat]

Ok, that didn't have the intended effect. I guess I don't understand where this recipe is specifying not to build for python 3.9. 😕

[xylar]

@rabernat, thanks for trying this. I'll fix it. And thanks for bringing this issue to our attention. I missed it in the last (few?) updates.

[xylar]

@conda-forge-admin, please rerender

[xylar]

Ok, that didn't have the intended effect. I guess I don't understand where this recipe is specifying not to build for python 3.9.

@rabernat, you have to rerender (as in the instructions above) in order for your changes to take effect. But by removing the skip entirely, you would have added windows and python 3.10 builds, neither of which will work. So the right thing to try is to bump up the skip to py >=310, as I have done.

We will see if the dependencies that gave us trouble before have loosened restrictions. Previously, there weren't python 3.9 builds on conda-forge of dependencies with the restrictions that Apache-Beam uses. It would likely be possible to go to their respective feedstocks and build new packages (including python 3.9) of those older packages. But often as a maintainer it's a rabbit hole that isn't worth the time unless you know it's needed.

[xylar]

Looks like we're in business!

[xylar]

@rabernat, give it an hour or so and the python 3.9 packages should be there.

[rabernat]

Xylar thank you so much for your help here!

It's great to run into an old friend on GitHub. Hope to reconnect some time in the near future. I just stalked you GitHub profile and it looks like you've been doing lots of cool data orchestration stuff! 🤓 I'd love to explain what we are up to with Pangeo Forge. If you're curious what we are using Beam for, the answer is here: pangeo-forge/pangeo-forge-recipes#256

[xylar]

Yeah, I just heard about this last week when @jbusecke gave a great seminar for our group at LANL! I didn't know that Apache Beam was part of the architecture, though. That's quite interesting.

I'm thinking of joining the pangeo biweekly meetings if that would be welcome. It would be fun to work toward more support for unstructured meshes.

[aaltay]

Thank you @xylar for keeping this updated. And hi @rabernat. I was not aware of pangeo-forge considering Beam. Happy to see that. And feel free to reach out if you think that we can help.

It would be amazing to see geoscience related talks at Beam summit or other Beam events, if your community would be interested in that :)