[2037]

[enssow]

Description

Fixes issue with selecting channels that don't use all available pressure levels or, a single level e.g. u_500 being treated as a surface variable.

The changes in the code will create a pressure index with all available levels but fill a speicifc variable at a given pressure level with Nans if not requested in hte code

e.g. only q_500 was requested to the ds["q"].values returns:

array([[[       nan],
        [       nan],
        [       nan],
        ...,
        [       nan],
        [       nan],
        [       nan]],

       [[0.00027466],
        [0.00028229],
        [0.00030136],
        ...,
        [0.00387573],
        [0.00354004],
        [0.00379944]]], shape=(2, 40320, 1), dtype=float32)

Closes two issues as refactored the find_pl to be better suited for the second issue as well

Issue Number

Closes #2037
Closes #2030

Is this PR a draft? Mark it as draft.

Checklist before asking for review

• I have performed a self-review of my code
• My changes comply with basic sanity checks:
  • I have fixed formatting issues with ./scripts/actions.sh lint
  • I have run unit tests with ./scripts/actions.sh unit-test
  • I have documented my code and I have updated the docstrings.
  • I have added unit tests, if relevant
• I have tried my changes with data and code:
  • I have run the integration tests with ./scripts/actions.sh integration-test
  • (bigger changes) I have run a full training and I have written in the comment the run_id(s): launch-slurm.py --time 60
  • (bigger changes and experiments) I have shared a hegdedoc in the github issue with all the configurations and runs for this experiments
• I have informed and aligned with people impacted by my change:
  • for config changes: the MatterMost channels and/or a design doc
  • for changes of dependencies: the MatterMost software development channel