Add option to exclude input cubes in output of multimodel statistics

[Peter9192]

Description

• Closes Multi model statistics: missing files for individual models  #975
• Link to documentation:

---

Before you get started

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

Checklist

• PR has a descriptive title for the changelog
• Code follows the style guide
• Documentation is available for new functionality
• Circle/CI tests pass
• Codacy code quality checks pass
• Documentation builds successfully on readthedocs
• Unit tests are available

---

To help with the number pull requests:

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

[bouweandela]

It would also be good to add a unit test to ensure that this doesn't get broken again in the future.

[stefsmeets]

If this is never going to be exposed and only applies to PreprocessorFile, why not fix it in the public function?

Suggested change

	keep_originals=keep_originals,
	statistics_products = _multiproduct_statistics(
	products=products,
	statistics=statistics,
	output_products=output_products,
	span=span,
	)
	if keep_originals:
	return products | statistics_products
	else:
	return statistics_products

[valeriupredoi]

@stefsmeets when you suggest code changes via the Github interactive editor there is no need to add text formatting syntax (like ```python here). Also @Peter9192 I would change keep_originals to something like save_input_cubes since originals refers more to the initial un-preprocced data and the option name doesn't tell the user what it does. Also note that for this to be in the documentation needs a mention of it 👍

[valeriupredoi]

OK I confirm that this fixes #975 - before and after fix:

(iris3) [valeriu@sci2 esmvaltool_var_test]$ ls -la eo1/recipe_python_20210202_112301/preproc/diagnostic1/ta/
total 18284
drwxr-xr-x 2 valeriu users       0 Feb  2 11:23 .
drwxr-xr-x 4 valeriu users       0 Feb  2 11:23 ..
-rw-r--r-- 1 valeriu users    1352 Feb  2 11:23 metadata.yml
-rw-r--r-- 1 valeriu users 9360338 Feb  2 11:23 MultiModelMean_Amon_ta_2000-2002.nc
-rw-r--r-- 1 valeriu users 9360338 Feb  2 11:23 MultiModelMedian_Amon_ta_2000-2002.nc
(iris3) [valeriu@sci2 esmvaltool_var_test]$ ls -la eo2/recipe_python_20210202_113151/preproc/diagnostic1/ta/
total 36584
drwxr-xr-x 2 valeriu users       0 Feb  2 11:32 .
drwxr-xr-x 4 valeriu users       0 Feb  2 11:32 ..
-rw-r--r-- 1 valeriu users 9370329 Feb  2 11:32 CMIP5_CanESM2_Amon_historical_r1i1p1_ta_2000-2002.nc
-rw-r--r-- 1 valeriu users 9367345 Feb  2 11:32 CMIP5_MPI-ESM-LR_Amon_historical_r1i1p1_ta_2000-2002.nc
-rw-r--r-- 1 valeriu users    2846 Feb  2 11:32 metadata.yml
-rw-r--r-- 1 valeriu users 9360338 Feb  2 11:32 MultiModelMean_Amon_ta_2000-2002.nc
-rw-r--r-- 1 valeriu users 9360338 Feb  2 11:32 MultiModelMedian_Amon_ta_2000-2002.nc

@Peter9192 cheers for the very quick fix, we need to get this in v2.2 pronto @jvegasbsc - we need a test this functionality is preserved as @stefsmeets says, please add it here 🍺

[Peter9192]

Here's a very simple test. I don't have any more time right now.

[valeriupredoi]

looks good to me, cheers @Peter9192 🍺

[LisaBock]

Works for me as well! Thanks!