Make sure climate statistics always returns original dtype

[zklaus]

Description

This adds code to the climate_statistics preprocessor to detect whether it has changed the dtype of the cube. If this is the case, it changes the dtype back to the original dtype.

Closes #1235

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
• 🛠 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 #1237 (9d341bc) into main (2a0bb11) will increase coverage by 0.00%.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##             main    #1237   +/-   ##
=======================================
  Coverage   85.54%   85.54%           
=======================================
  Files         188      188           
  Lines        9150     9154    +4     
=======================================
+ Hits         7827     7831    +4     
  Misses       1323     1323           

Impacted Files	Coverage Δ
esmvalcore/preprocessor/_time.py	94.21% <100.00%> (+0.07%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 2a0bb11...9d341bc. Read the comment docs.

[zklaus]

This line looks a bit fishy to me, but it was there before, so there is that. It looks like this should be clim_cube.remove_coord(clim_coord). @valeriupredoi, this seems to come from you. Do you have an opinion?

[jvegreg]

I am changing the tests in the time module to always check for the types

[jvegreg]

I have detected the same problem when computing standardized anomalies, as they get converted to float64 from a float32 input. I am trying to solve it and then I will push the changes

[zklaus]

I have detected the same problem when computing standardized anomalies, as they get converted to float64 from a float32 input. I am trying to solve it and then I will push the changes

Be advised that I came to this solution via the anomalies. Turns out the problem with anomalies was that the climate_statistics, which are used in the calculation of the anomalies did this conversion, so maybe once this PR is in, the anomalies problem vanishes.

[jvegreg]

Be advised that I came to this solution via the anomalies. Turns out the problem with anomalies was that the climate_statistics, which are used in the calculation of the anomalies did this conversion, so maybe once this PR is in, the anomalies problem vanishes.

I am working directly on this PR, so that's not the issue

[jvegreg]

Nice catch! I currently do not have access to data (BSC annual maintenace) to properly test with a recipe, though.