Standardise variable name for clearsky GHI and DNI, add to glossary

[RDaxini]

Is your feature request related to a problem? Please describe.
I have counted three variations of clear sky GHI in the docs:

clearsky (in clearsky.detect_clearsky)
clearsky_ghi (in irradiance.clearsky_index)
ghi_clearsky (in irradiance.dirindex)
and no entry in the variables and symbols page (soon to become a glossary #2234)

for DNI, we have dni_clear in the variables and symbols page but also clearsky_dni (in irradiance.dni) and dni_clearsky (in irradiance.dirindex) in the docs.

Describe the solution you'd like

• Agree a single format
• Add/modify an entry in the variables and symbols page

Additional context
Related: #1253

[RDaxini]

Except for just clearsky, all options seem reasonable to me. I think either clearsky_xxx (written as you'd say it) or xxx_clear (shorter) would be best.

[cwhanse]

I prefer ghi_clear but it's a mild preference. Parallels the few terms that are a conditional irradiance, i.e., dni_clear, ghi_extra, poa_global, etc.

[RDaxini]

I think we're set on ghi_clear (also dni_clear for where it isn't already used)

Would updating these variables be classed as breaking changes and require some kind of deprecation? I am thinking about #2237
Is a separate PR required for each function that is changed, or is one PR for all the functions involving the changing variable appropriate? (e.g. #2273)

[adriesse]

Did you consider having just one PR for these two names since the changes are so closely related?

[RDaxini]

Did you consider having just one PR for these two names since the changes are so closely related?

@adriesse I asked about PR structure here but thought each variable had to be separate, no response so I was not really sure. I confess I think I underestimated the overlap. Is it possible to combine PRs after creation? We should do that if it would help with merging later.