Unpin pandas

[valeriupredoi]

Description

Continual modernization of our environment, see #3812
Needs ESMValCore=2.12 stable release (can test with RCX if need be)

---

Before you get started

• ☝ Create an issue to discuss what you are going to do

Checklist

It is the responsibility of the author to make sure the pull request is ready to review. The icons indicate whether the item will be subject to the 🛠 Technical or 🧪 Scientific review.

• 🛠 This pull request has a descriptive title
• 🛠 Any changed dependencies have been added or removed correctly
• 🛠 All checks below this pull request were successful

[bouweandela]

Has #3674 been solved?

[valeriupredoi]

Has #3674 been solved?

Just as in Ford v Ferrari when Shelby said take her to 7000+ RPM - one way to find out 😁

[sloosvel]

Hi, the code freeze for version 2.12 will be on Monday 3rd of March. Is this ready to be included? If this is not ready, let me know and I will move it to the next milestone.

[valeriupredoi]

yeah! This needs be plopped in 2.12 - @bouweandela pls approve 🍺

[bouweandela]

If we include this in v2.12, @sloosvel will have to test all the R recipes again. Would you be happy to do that @sloosvel?

[bouweandela]

Even better, test all the R recipes with this branch before merging because I'm not convinced that #3674 has been solved.

[sloosvel]

Ok, I can have a go at testing the R recipes. I guess I should create a new environment right?

[valeriupredoi]

true hero @sloosvel 🎖️ Yeah just the r-base needs be unpinned but I guess we can use the bot here instead of you running them recipes? Well, try a couple, if they need special data the bot doesn't have, then manual rerun should be done

[sloosvel]

I have to admit I am not quite sure how to run the bot 😅 Some of the recipes require quite a big amount of resources anyway, three of the R recipes never finished running during the 2.12 testing despite increasing the time limit of the jobs... For the moment I am creating a new environment which is taking forever to solve.

[valeriupredoi]

just as the Circle test https://app.circleci.com/pipelines/github/ESMValGroup/ESMValTool/13797/workflows/21f50cc0-95cc-4959-9731-80f01f87b1c6/jobs/80790 - the problem is that unpinning these two may actually be very harsh on our already hard to solve env - let me pop in a GHA see the times

[sloosvel]

It's been two hours and it's still solving for me...

[valeriupredoi]

yeah that's not promising at all, lemme have a look at what the GHA does

[valeriupredoi]

@sloosvel the GHA tests run fine and envs get solves in 2-4min - make sure you are using mamba=2+, otherwise with an older mamba things get veeery slow

[sloosvel]

I upgraded mamba:

[b381943@levante1 ~]$ mamba --version
2.0.5

But it still gets stuck solving the environment, the loading bar does not even move. I'll try over the weekend and let's see what happens

[valeriupredoi]

don't make it a priority, Saskia, it's good to have the pin on pandas removed, but the one on r-base doesn't matter that much, it's not a massive blocker against anything. The big blocker is NCL anyway. Have a good weekend, don't work 🍺

[sloosvel]

The following R recipes fail using this branch:

• recipe_combined_indices.yml
• recipe_extreme_events.yml
• recipe_miles_block.yml
• recipe_miles_eof.yml
• recipe_miles_regimes.yml
• recipe_quantilebias.yml

[valeriupredoi]

Thanks, Saskia! Is there any R recipe that runs fine? 🤣
Ok, let's keep the r-base pin for now then, and only unpin 🐼