EvalML grib exporter

[kctezcan]

Description

A new PR for adding a grib exporter. I closed the other PR and started from a fresh branch because refatoring/modifying the other one was more work.

We need to export to grib to able able to use the internal MeteoSwiss evaluation tool (https://github.com/MeteoSwiss/evalml). Claude wrote an exporter for this purpose

Tested the code, so now we can export to grib and use evalML to generate scores, below the scores for an ICON 1km 1hr model vs measurement stations in Switzerland (https://peakweather.readthedocs.io/latest/) (scores are bad as the model has not been trained for long):

Issue Number

No issue

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

[kctezcan]

@clessig I closed the other PR and started fresh. Have a look here please, this should have more minimal changes. Thanks

[clessig]

@enssow : Could you have a look at the PR and pull in other expertise as needed (there is a change to io.py that maybe @grassesi needs to answer).