Cmorize eobs

[mwjury]

@mattiarighi here for E-OBS tas, tasmin, tasmax, pr, psl.
Again with additional monthly means.
I had to fix fix_coords in utilities.py so it can handle asymmetrical longitudes (i.e. not centered at zero).

[mattiarighi]

I'm getting the data.
Could you have a look at the conflicts? Thanks! 👍

[mwjury]

@mattiarighi The conflict is due to a inconsistency in fix_coordinates, which in its old form can not handle asymmetrical longitudes (i.e. not centered at zero).

[mattiarighi]

There is still a conflit in utilities.py, git cannot merge the branch if this is not solved manually.

[mattiarighi]

@valeriupredoi I need your review on the changes in utilities.py.

[mattiarighi]

@valeriupredoi I tested a few datasets to make sure the changes in utilities.py are back-compatible.
Can you have a look anyway?

[mattiarighi]

@valeriupredoi can I have an approval please! 🍻

[valeriupredoi]

One sec, on it, a bit late in work today - trying to catch a slightly emptier bus to work 🍺

[valeriupredoi]

I reviewed only the technical bits I assume @mattiarighi you've seen the data provenance stuff. Please have a look at the two important comments that transform longitude points, the other two bits are pythonic stuff 🍺

[valeriupredoi]

Suggested change

	for res in cfg['attributes']['resolution'].values():
	for _, res in cfg['attributes']['resolution'].items():

[bouweandela]

Why would you want to change this? It looks fine as it is?

[bouweandela]

@valeriupredoi I noticed that Codacy is complaining about importing private functions (i.e. starting with _) from the utilities.py. Would it be possible to make those functions public in another pull request? Because I guess they're supposed to be used by the cmorizers, right?

[valeriupredoi]

funny, this is not like accessing a protected member from outside the class, but meh, yes I'll do it! 🍺

[valeriupredoi]

actually - can we skip it and tell Codacy to go hike? It's an absolute monster to change all the cmorizer scripts, plus, I don't feel like writing documentation for each of these funcs and publishing it, it's something that it's prob best to stay private. What do you think @bouweandela

[bouweandela]

I agree about not writing the documentation, that would be too much work for very few users, because these functions are only used by people writing cmorizer scripts. However, I think that renaming them without the leading underscore shoudn't be that much work, just use something like

find esmvaltool/cmorizers/ -name '*.py' -exec sed -i 's/_fix_bounds/fix_bounds/' {} +

and carefully look at the git diff before committing to make sure you didn't accidentally change other stuff.

[bouweandela]

@mattiarighi Looks good to me, can you please test and merge?