More extract features for RDRS

[juliettelavoie]

Pull Request Checklist:

• This PR addresses an already opened issue (for bug fixes / features)
  • This PR fixes Extract Features that would be nice with RDRS #180
• (If applicable) Documentation has been added / updated (for bug fixes / features)
• This PR does not seem to break the templates.
• HISTORY.rst has been updated (with summary of main changes)
  • Link to issue (:issue:number) and pull request (:pull:number) has been added

What kind of change does this PR introduce?

• These features are in the ESPO-G workflow but I would like to have them in xscen to clean up the workflow. They were added because of RDRS.
• feature mask: Either pass a dataset of a mask or find the mask in the catalog (need to do that in search_data_catalog). The mask is used to add nans to the data.
• new region method sel: Instead of cutting on lat-lon (for ERA5-Land), I want to cut in rlat-rlon (for RDRS). I want this to be only defined in the config. I know it is a bit weird to put a non-clisops method in the clisops_subset function, but I also call this function in my workflow and I need it to cut the region regardless of if it is defined in rlat-rlon or lat-lon. Would it be too much of a breaking change to change the name of the function ?

Does this PR introduce a breaking change?

not, if we don't change the name of clisops_subset

Other information:

[aulemahal]

I agree that clisops_subset is a bit misnamed. I would support a renaming, and may be a move to xscen.spatial ?

We could lower the breakingness of this change by keeping a clisops_wrapper that emits a FutureWarning for one version.

[aulemahal]

Could we also accept a DataArray ?

I don't find the "bool" option very useful, but I see that #183 would complete this feature.

[juliettelavoie]

sure for DataArray.
I guess it is only useful for me now that I already have a mask that is in the catalogue.

[RondeauG]

I just saw that clisops_subset still exists as is. I think that we should rather have:

def clisops_subset(ds, region):
     warning(blabla)
     ds_subset = xs.spatial.subset(ds, region)
     return ds_subset

So as not to duplicate the code twice and have an outdated version, especially since the new function should work perfectly well with the old arguments.

[juliettelavoie]

I just saw that clisops_subset still exists as is. I think that we should rather have:

def clisops_subset(ds, region):
     warning(blabla)
     ds_subset = xs.spatial.subset(ds, region)
     return ds_subset

So as not to duplicate the code twice and have an outdated version, especially since the new function should work perfectly well with the old arguments.

done!

[juliettelavoie]

In extract_dataset, we ask for the name in the definition of the region in order to update cat:domain. This field was not necessary in clisops_subset, but I made it necessary in spatial.subset.