Add theory error to matching grids

[giacomomagni]

Include the theory error in the matching grids using the 7 points prescription

[Radonirinaunimi]

This LGTM! The uncertainties look better now. I will run a fit and hopefully the results will look better.

[Radonirinaunimi]

The report for the fit including theory uncertainties is available here: https://data.nnpdf.science/NNUSF/reports/theory-unc-ac1b7cd-221009/output/

The results look good. After other people go through the report, then I can merge this.

[juanrojochacon]

Looks good @Radonirinaunimi ! Just a small comment, I thought we agreed to fluctuate also the pseudo-data, but perhaps this is done on a separate branch and not on this one?

[Radonirinaunimi]

Yes, indeed, this is still done in a separate branch #56, mainly because we also wanted to assess the effects of each change separately.

[juanrojochacon]

Which is always a very good practice so I don't complain ;)

[Radonirinaunimi]

Merging this as it is resolved. The last fit with replica 0 as CV and $\chi^{2, \rm real}_{\rm exp}$ fixed is available at https://data.nnpdf.science/NNUSF/reports/fixexpr-f9563d5-221011/output/