Cleaning in yamls and sources

[benjaminmenetrier]

I think some cleaning might be required in yamls: there are a lot of redundancies, for instance the variables in https://github.com/andreapiacentini/aq/blob/Feature/ChangeVar/test/testinput/dirac_loc_4d.yaml.

And also in the source. I saw that the QG model renamed:

• ObsAuxControl into ObsBias
• other ObsAux[...] into ObsBias[...]

I think it would be useful to stay close to the QG model in order to make updates easier.

[andreapiacentini]

Thanks Ben,
I'll have a look to the yamls redundancies. Some of the 4d yamls are generated by a python script that is not very keen in using mnemonic shorcuts.

Emanuele is our "Obs" routines master. I'll let him modify at best the ObsAux routines (yet ObsAux is the naming used in oops, isn't it?) toward the corresponding Qg choice, even if all the ObsBias treatments are probably not of immediate interest for AQ.

The updates are an open issue for us: what's the workflow to stay in phase with the fast evolving oops develop branch?
Time has come for a new brainstorming on localisations and on these workflow choices. Will you be around in a relatively near future?

[benjaminmenetrier]

You are right, they are called ObsAux in the generic part of OOPS, so we might as well keep them that way in AQ. My agenda is open this week: we can talk any time except tomorrow (Tuesday).

[andreapiacentini]

I modified the yamls and spotted a wrong obs error specification in NO_2. The reference changes as well.
Even if we decide to keep the ObsAux naming conventions, I'd like if Emanuele could go through it, clean some misporting (e.g. the hardwired ntypes=4) and add some warning print statement (or ABORT) if anything were calling them since we did not really tested it.
Suggested new meeting on Thursday at 3PM CET