Update theory documentation

[felixhekhorn]

We need to update the theory documentation:

• [ ] index.rst
• [ ] FastInterface.rst
• [ ] PTevol.rst
• [ ] theoryindex.rst
• [ ] theoryparamsdefinitions.rst
• [ ] theoryparamsinfo.rst

FastInterface.rst

We should replace the old documentation here https://docs.nnpdf.science/theory/FastInterface.html as it is outdated. Instead the page should contain the updated way:

• to calculate FK tables we need two (main) programs: pinefarm + pineko
• the mathematical concept of FK tables (with updated notation) is here: https://pineko.readthedocs.io/en/latest/theory/fktables.html
• the actual computation is a two step process:
  1. generate a (process-dependent) grid with pinefarm
  2. transform the grid into an FK table
• for 1.
  • point to https://pinefarm.readthedocs.io/en/latest/index.html
  • point to https://github.com/NNPDF/pinecards
• for 2: point to https://pineko.readthedocs.io/en/latest/index.html

[felixhekhorn]

Questions for @scarlehoff :

1. I'd like to drop the explicit table in theoryindex.rst, fine?
2. is vp-list -r theories still working? is anybody using that?
3. is vp-checktheory still working? is anybody using that?
4. is validphys.loader.Loader.check_theoryID stil working? is anybody using that?

[scarlehoff]

1. Mmm, that one is painless and autogenerated, I think it is fine.
2. It does work, not sure how many people use vp-list in general, but it does work.
3. Same. I don't use it, not sure others.
4. Yes. It is what vp uses internally to download theories.

2 and 3 will work as long as the theories work because there's no really much functionality in those scripts, it is just a handy way of reading a theory and printing it

[RoyStegeman]

I use neither of these commands.

vp-list seems quite useless since the list is too long to read "manually" and I can't really many people grepping to see if a theoryID exist.

vp-checktheory doesn't even fit the full Comments, so again, not sure how useful that really is. Also, as soon as you want to do something as trivial as comparing two theories you already need something else.

I'd be in favour of removing them. If people actually rely on them we'll hear the complaint

[felixhekhorn]

re 1: How useful is it to have a table with 52 columns and 643 rows? on my screen it takes ~5500 x 21000 pixels ... yes I have used the page before myself, but still I wonder ... I'm worried about the power set of PTO x MHOU x QED x mt x alphas x ...

[scarlehoff]

This is actually something I was wondering also yesterday when we discused about 41_001_000, maybe we should only printout the _000 theories, for which the comment should explain to what category do they belong

So we would have

41_000_000 standard 4.1 settings 41_001_000 4.1 grids 4.0 settings 41_100_000 polarized

etc

And then you know in which subgroup you have to look for the theory you want?

(and, as @enocera suggested, remove all old theories)

(or we can just remove the table and everyone looks in the folder every time, now it is no longer a binary file so no need to download it like before)