Feat/awicm3.2

[JanStreffing]

This merge brings AWI-CM3 v3.2 (https://awi-cm3-documentation.readthedocs.io/en/latest/releases.html#awi-cm3-v3-2) to the esm_tools release branch

[mandresm]

Hi @JanStreffing, nice PR!

I have some comments about information duplication in awicm3.yaml. Other than that there are a bunch of new namelists in namelists/fesom2/2.0/awicm3/v3.2 and another set in namelists/fesom2/2.5/awicm3/v3.2. Is this intentional? I would think that you just wanted the ones of fesom 2.5.

[JanStreffing]

Question:
I'm trying to use the new fesom2.5.yaml. like this:

in awicm3.yaml:

fesom:
        version: "2.0"
        choose_general.major_version:
                "v3.2":
                        add_version: "2.5"

Is this the right way? In the finished yaml I still get:

    debug_info:
        loaded_from_file: /home/a/a270092/esm_tools/configs//components/fesom/fesom-2.0.yaml
    destination: fesom-2.0

[mandresm]

Question: I'm trying to use the new fesom2.5.yaml. like this:

in awicm3.yaml:

fesom:
        version: "2.0"
        choose_general.major_version:
                "v3.2":
                        add_version: "2.5"

Is this the right way? In the finished yaml I still get:

    debug_info:
        loaded_from_file: /home/a/a270092/esm_tools/configs//components/fesom/fesom-2.0.yaml
    destination: fesom-2.0

I don't think so, in principle, it should be possible to run it without the add, but there is a problem there, and it's being there for a while (sorry...). What I've tipically done in this case is to include all the versions in the choose and not define one default one. But this might not work either, at least for esm_master. I'll try to have a look at a solution this afternoon.

[mandresm]

I can confirm that there is a bug here. I am half way of finding out what's going on, but I'm in a conference, so I'll be slower than usual.

[JanStreffing]

Thanks for digging!

[mandresm]

Hi @JanStreffing,

I have managed to fix the problems with the version switching... the functions affected by the bugs were written by me on my early days of ESM-Tools :)

You'll find the fixes in the branch fix/recursive_resolution_of_early_variables_and_fix_find_key. You can just merge that branch here (just departed from release) or cherry-pick the single commit 3028599d2a27fdec65cba2e4c0f498eaafdec544 because I don't think the PR (#1063) will be reviewed earlier than Wednesday.

The failed tests in #1063 shouldn't be a problem.

[JanStreffing]

ready to merge 🍻

[JanStreffing]

Halt! one more fix...

[JanStreffing]

done

[mandresm]

Are you aware of this changes in the namcouple of the frontiers version?

https://github.com/esm-tools/esm_tools/actions/runs/6572115118/job/17852600385?pr=1056#step:8:395

[JanStreffing]

I was not! I'll fix that.

[mandresm]

I'm adding a test runscript to the CI for the 3.2 version and doing a real test in Levante. Should I perform the real test somwhere else? (that won't delay the merge, and I'll do it between today and tomorrow)

[JanStreffing]

i think it's a good idea to do a test on juwels as well. We might not have albedo yet, correct?

[mandresm]

i think it's a good idea to do a test on juwels as well. We might not have albedo yet, correct?

Albedo is something for next week. I'm running the tests on Levante. I'll do the same on Juwels now that all is configured in esm_tests. Tomorrow I can do the same in Aleph.

[JanStreffing]

great 👍 :

[mandresm]

#approve-changes

[mandresm]

#bump