Implement PSP8 Pseudopotential format

[mfherbst]

The PSP8 pseudopotential format seems to be worth supporting for experimentation.

[antoine-levitt]

There are a few formats around. Any format for which we can get good pseudos (ONCVPSP for physical accuracy, and Goedecker's for mathematical niceness) is good, so it comes down to difficulty of writing a parser. Also it's probable somebody wrote a parser in a sane language (eg python) from which we can maybe steal.

[antoine-levitt]

Just looked into this, it doesn't look that bad. Additional complications (on top of the usual "figuring out what they wrote in the 'spec'") wrt what we have is a possible core correction (to be checked, but I think it's just a radial density that needs to be added to the pseudo-density in the exchange-correlation term), and numerical potentials/projectors for which we'll need to do quadrature.

[mfherbst]

I agree. I looked at them a while back, but then what made me drop it was the bad spec. ONCVPSP are better specified ...

[azadoks]

Should follow easily from work done in #741 . Once NLCCs are implemented, the only other feature of PSP8 that's missing is spin-orbit coupling.

[antoine-levitt]

Do we need psp8 support? Can't we use external converters from psp8 to upf?

[azadoks]

I'm not sure if one exists, but I think the most used psp8 pseudos are those from PseudoDojo which are also available as UPFs (not converted, just also output by ONCVPSP afaik).