Clustering of chi2 values per process (not per experiment) in comparefits

Created by: juanrojochacon

One feature of the validphys report that would be really useful to implement is that the chi2 is clustered by type of process, rather than by experiment (or better said, this would be a new table in addition of the one we have now). The reason is that the discussions in the paper are easier if one refers to the "chi2 of DY" or "chi2 of jets", and this feature needs to be added by hand each time we want it. For example @siranipour we made use to this feature quite extensively in the strangeness paper.

Would it be possible to add it to the default report? I think we really need it to write the NNPDF4.0 paper.

Thanks!