Implement first prototype to be tested

[sophie-xhonneux]

Description

When switching between ERA5 and Operational analysis we can select the same channels but they may have a different ordering making it hard to re-use pretrained encoders...

Issue Number

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

[clessig]

This should be handled in the data readers and not in the trainer. The channels are available there and we only need to ensure that there's a stable ordering when the channel data is passed to the model. See

WeatherGenerator/src/weathergen/datasets/data_reader_anemoi.py

Line 110 in d30421f

self.source_idx = self.select_channels(ds, "source")

and the related lines below in the file.