Will this package work if we remove `**_` from all function signatures?

[khsrali]

@mikibonacci @GeigerJ2 @npaulish
Discussed during the coding days, unfortunately with having this we cannot throw an error if an invalid input is provided.
aiidateam/team-compass#25

[mbercx]

I just checked, and I'm pretty sure we could in the get_builder_from_protocol signatures. IIRC, the reason these are here is to be able to pass inputs in the kwargs of higher-level work chains to lower-level ones without things breaking in case there are different inputs to two different sub work chains. Within aiida-quantumespresso, I'm pretty sure we could avoid this.

I never liked the **_, nor the kwargs. Having a typo in your input keyword just pass without issue isn't great UX, nor is the opaque kwags that you'll find in a lot of AiiDA constructors/methods. So happy to fix that here. The problem is that it's technically backwards-incompatible: other packages building on top of aiida-quantumespresso may rely on it. So we'd have to think of a deprecation pathway and save this for v6.0.

Re the ones in hubbard.py, I'm not sure. Will ping @bastonero for his feedback.

[bastonero]

Hello everyone, in all the packages I developed I also always used **_. It was very useful to pass, e.g., electronic_type, spin_type keywords much more easily. Don't know if this will eventually also affect the workchains in aiida-hubbard, etc.

To be tested?