total chi2/phi use experiments

Created by: wilsonmr

provided no theory covmat is being used. I don't see an imminent issue with this and it significantly speeds up comp fits

@tgiani please can you run a compfits using this and check the numbers are the same in particular in the fit summary table and the total row of the chi2 table?

Note I also changed the provider name for some plot functions which I guess we never use because the providers were non-existent before. Hopefully they work now.

Finally I also added a small fix for a bug I guess was introduced in #862 where a function relied on an import of linalg but the import didn't exist @RosalynLP can you check you agree with that? (EDIT: specifically in results.py with experiments_invcovmat)

closes #860 (closed) closes #924 (closed)