Allowed ignoring scalar coordinates in multi_model_statistics

[schlunma]

Description

This PR allows the optional keyword ignore_scalar_coords for the multi_model_statistics and ensemble_statistics preprocessors. With this setting enabled, different scalar coords in the input cubes are ignored and do not raise merge MergeErrors.

Closes #1606

Link to documentation: https://esmvaltool--1934.org.readthedocs.build/projects/ESMValCore/en/1934/recipe/preprocessor.html#multi-model-statistics

---

Before you get started

• ☝ Create an issue to discuss what you are going to do

Checklist

It is the responsibility of the author to make sure the pull request is ready to review. The icons indicate whether the item will be subject to the 🛠 Technical or 🧪 Scientific review.

• 🧪 The new functionality is relevant and scientifically sound
• 🛠 This pull request has a descriptive title and labels
• 🛠 Code is written according to the code quality guidelines
• 🧪 and 🛠 Documentation is available
• 🛠 Unit tests have been added
• 🛠 Changes are backward compatible
• 🛠 Any changed dependencies have been added or removed correctly
• 🛠 The list of authors is up to date
• 🛠 All checks below this pull request were successful

---

To help with the number pull requests:

• 🙏 We kindly ask you to review two other open pull requests in this repository

[codecov]

Codecov Report

Merging #1934 (6c587dd) into main (c25b930) will increase coverage by 0.00%.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##             main    #1934   +/-   ##
=======================================
  Coverage   92.18%   92.18%           
=======================================
  Files         234      234           
  Lines       12205    12213    +8     
=======================================
+ Hits        11251    11259    +8     
  Misses        954      954           

Impacted Files	Coverage Δ
esmvalcore/_recipe/check.py	91.26% <100.00%> (+0.27%)	⬆️
esmvalcore/preprocessor/_multimodel.py	97.88% <100.00%> (+<0.01%)	⬆️

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

[valeriupredoi]

very cool, Manu! Just a couple of very minor review points from me 🍺 Would we want to extend that [p0, ptop] list of coords to be always removed? If so, I'd put that at the top of the mm script, for now, maybe we'll take it out from there and expose it to the API some time in the future. About boolean (true or false) - it's always nice to tell the users stuff in layman terms, you know 😁

[schlunma]

Thanks for the quick review @valeriupredoi!

Would we want to extend that [p0, ptop] list of coords to be always removed? If so, I'd put that at the top of the mm script, for now, maybe we'll take it out from there and expose it to the API some time in the future.

I don't think we should expose this to the public API, especially with that new option now. We can also extend that list in there 👍

[valeriupredoi]

Thanks for the quick review @valeriupredoi!

Would we want to extend that [p0, ptop] list of coords to be always removed? If so, I'd put that at the top of the mm script, for now, maybe we'll take it out from there and expose it to the API some time in the future.

I don't think we should expose this to the public API, especially with that new option now. We can also extend that list in there +1

Perfect then! Even less knobs and buttons for users to deal with it - I like it. Great work on this, hope signor Lovato is 'appy now, after we've left him out for dry for such a long time 😁

[tomaslovato]

thanks a lot for the quick response @schlunma & @valeriupredoi

[valeriupredoi]

@ESMValGroup/technical-lead-development-team this is niice (in Borat accent) and ready for merging - please have a look and merge at earliest convenience, if you dillydally about it, myself or @remi-kazeroni will merge (given tight schedule on 2.8 freeze)

[zklaus]

It's a little bit muddled, isn't it? The documentation states

To ignore different scalar coordinates in the input datasets, [...]

and

Differing scalar coordinates can be ignored with the option ignore_scalar_coords.

but it's not quite clear what "ignore" really means and the code just removes all scalar coordinates.
This is stated correctly down in the notes on the API docs, where it says

If True, remove any scalar coordinate in the input datasets before merging the input cubes into the multi-dataset cube. [...]

In lieu of implementing the behavior suggested by the documentation, i.e. removing only the differing scalar coordinates, I suggest clarifying the documentation in the two places mentioned above to make it easier to understand for the user what is happening.

[valeriupredoi]

bit of a 🐔 - 🥚 problem, @zklaus -> ignoring stuff in input cubes means said stuff is not present in output cube(s) - that sounds very logical to me as @schlunma wrote it, but I would indeed change:

To ignore different scalar coordinates in the input datasets, use the option
ignore_scalar_coords: true.

with

To ignore all scalar coordinates in the input datasets, use the option
ignore_scalar_coords: true; this way these scalar coordinates will not be saved to the output multimodel cube.

[schlunma]

@zklaus I tried to improve the documentation, could you take another look? Thanks!