Update interpolate action to output HEALPix_nested directly

[tweska]

This change removes the need to use the renumber-healpix action to interpolate to a HEALPix_nested. You can now do it in the interpolate action directly.

Draft because I think we might want to consider adding a few more things:

• Extra 1024 test, I think the data needed already exists so we might as well.
• Tests in the test/action/interpolate directory which will remain after we remove renumber-healpix
• Some sort of warning message in the renumber-healpix, alerting users of this change.

[pmaciel]

I would remove the renumbering option altogether, as well as associated code -- this is legacy and not needed.

[MircoValentiniECMWF]

Removing renumber HEALPix requires a lot of modifications in climateDT. RAPS, multIO plans, and moreover there is still the FESOM interpolation (interpolation from matrix) pending question (who will remap the matrices?). My understanding is that now we are in a very busy phase and it is better to not touch for the moment.

[pmaciel]

Again I'd remove all cases/code refering to renumbering

[pmaciel]

Excellent -- though I'd remove teh renumebering functionality/tests

[FussyDuck]

All committers have signed the CLA.

[tweska]

Added a test test with reference data generated without the changes in the interpolate action. The output is exactly the same as before, even if you use the original plan. As requested by @stefaniereuter

The commit from @MircoValentiniECMWF is a tool which lets you convert the matrices needed for fesom, this is work in progress. We still need to test it and possibly make a few changes.

[MircoValentiniECMWF]

some more points:

1. It will be nice to add to this script "src/multio/action/interpolate/generate_fesom_caches.sh" the management of the new command line flag to reorder the matrix ( or to create two scripts generate_fesom_cache_ring.sh generate_fesom_cache_nested.sh)
2. @dsarmany what shall we do with the HEALPix.* class? now I just copied the files in the interpolate action, but we have a copy in the renumber HEALPix action. Maybe we should put this calss in utils? or is it ok to have a duplication since renumberHELAPix should be removed in the future?
3. We need to update the interpolate action also for the "matrix" interpolation method. Same modification we made for the other method we need to inject the ordering convention (that has to be in the matrix name.)
4. It will be nice to have a test for the FESOM workflow. csv -> mat and then using the mat in the interpolate action to do an interpolation from FESOM grid to HEALPix nested directly

[pmaciel]

some more points:

1. It will be nice to add to this script "src/multio/action/interpolate/generate_fesom_caches.sh" the management of the new command line flag to reorder the matrix ( or to create two scripts generate_fesom_cache_ring.sh generate_fesom_cache_nested.sh)

2. @dsarmany what shall we do with the HEALPix.* class? now I just copied the files in the interpolate action, but we have a copy in the renumber HEALPix action. Maybe we should put this calss in utils? or is it ok to have a duplication since renumberHELAPix should be removed in the future?

3. We need to update the interpolate action also for the "matrix" interpolation method. Same modification we made for the other method we need to inject the ordering convention (that has to be in the matrix name.)

4. It will be nice to have a test for the FESOM workflow. csv -> mat and then using the mat in the interpolate action to do an interpolation from FESOM grid to HEALPix nested directly

For one of the points above, you can also modify the mir-codec-to-weight-matrix tool to renumber the input and/or output indices directly.

[dsarmany]

2. @dsarmany what shall we do with the HEALPix.* class? now I just copied the files in the interpolate action, but we have a copy in the renumber HEALPix action. Maybe we should put this calss in utils? or is it ok to have a duplication since renumberHELAPix should be removed in the future?

I think it is okay to keep the duplicates, given that we want to get rid of the whole action anyway. I am happy to give climate DT a bit of a breathing space right now. But at some point, and still sooner rather than later, I agree with @pmaciel that we should just simply remove this action and force the climate-DT plans to be updated.