fix: multiple datasets utility

[JPXKQX]

Description

What problem does this change solve?

What issue or task does this change relate to?

Additional notes

As a contributor to the Anemoi framework, please ensure that your changes include unit tests, updates to any affected dependencies and documentation, and have been tested in a parallel setting (i.e., with multiple GPUs). As a reviewer, you are also responsible for verifying these aspects and requesting changes if they are not adequately addressed. For guidelines about those please refer to https://anemoi.readthedocs.io/en/latest/

By opening this pull request, I affirm that all authors agree to the Contributor License Agreement.

[VeraChristina]

Thanks for moving this! Just two points:

• Do we also need get_dataset_data_config? For the import Ana found in models/src/anemoi/models/interface/init.py
• Is there a better place than utils given that it's somewhat specific to how the training config files are set up? (I understand that it needs to be in a dependency of models, so it can't go in training itself, and perhaps the best solution for now is utils, just wondering) cc @gmertes
  edit: how about putting it into models? that's also not ideal, but it's used in models and training, and so it would remain in core?

[JPXKQX]

• The get_dataset_data_config is deleted now.
• I could see this function being used for inference, too, to support the same config structure. What are the plans?

[VeraChristina]

Looks like it's not needed in inference as the approach to handling the different config layouts is via two different subclasses of the Metadata class (which handles this more explicitly, storing relevant metadata entries in cached properties). So I'd have a preference to put it into models