Smarter removal of coordinate metadata in multi_model_statistics preprocessor

[schlunma]

Description

This PR makes the removal of coordinate metadata in multi_model_statistics preprocessor smarter, i.e., it uses common names (var_name, standard_name, long_name) for coordinates whose name() and units match.

Prior to that change, the code removed the long_name attribute for all coordinates that did not exactly match, which caused problems for coordinates that only had a long name (e.g., the ones coming out of iris.coord_categorisation).

Note that this does not apply if two or more coordinates of a cube have the same name().

In addition, now also the circular attribute is removed for all dimensional coordinates (this may be set by some models and by some not, which prevented multi_model_statistics to run).

Case	Old logic	New logic
Exactly identical coordinates	Remain unchanged	Remain unchanged
Coordinates with matching name() and units	Delete long_name and attributes	Set common var_name, standard_name, long_name, delete attributes, and set circular=False for dimensional coordinates
All remaining coordinates	Delete long_name and attributes	Delete long_name and attributes, and set circular=False for dimensional coordinates

Closes #1807

Link to documentation:

---

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 #1813 (c54dc58) into main (d832c8c) will increase coverage by 0.02%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##             main    #1813      +/-   ##
==========================================
+ Coverage   92.03%   92.05%   +0.02%     
==========================================
  Files         234      234              
  Lines       12042    12074      +32     
==========================================
+ Hits        11083    11115      +32     
  Misses        959      959              

Impacted Files	Coverage Δ
esmvalcore/preprocessor/_multimodel.py	97.70% <100.00%> (+0.32%)	⬆️

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

[valeriupredoi]

nice indeed, bar one (mostly philosophical) question: are we looking for full uniformity in metadata, or we can do something smart like: if 90% of datasets have a certain value for an attribute, then we'll use that (see inline comment specific to this question)

[valeriupredoi]

@ESMValGroup/technical-lead-development-team can one of you lovlies have a last look and merge this please? 🍺