Separate handling of groups and variables in timeseries (optional)

[Dooruk]

⚠️ I messed up and I already pushed my changes to develop, see the last two commits. It is a minor change and hopefully won't impact others but I learned my lesson. It is backward compatible.

This simple change allows you to use either the original configuration format or the new one with separate groups and variables. Also adds an example for insitu argo data.

Reduces redundancy IMO, you no longer need to repeat group names in your configuration.

With the new approach it is possible to define groups and variables separately, similar to how it is in other classes.

time_series:

  - begin_date: ...
    final_date: ...
    interval: PT6H

    collection: experiment
    groups:
      - omanPassedQc
      - ombgPassedQc
    variables:
      - salinity
      - waterTemperature
    aggregation_methods:
      - mean
      - count
      - std
    dimension:  Location

closes #231 and #239

[Dooruk]

Hmm. Looks like github is using python3.13 now and that might be messing up the requirements and DAG?

[karpob]

Looks fine to me.

[CoryMartin-NOAA]

The github CI drives me up the wall with this constant chasing bugs. I vote we force merge this and then try to fix the CI "soon"