Another ugly solution to using th_covmat in validphys report/fit summary table
Created by: wilsonmr
Title says it all.
especially having to remove the lru_cache seems unfavourable but this allows the user to change the covmat to include theory contribution at the level of the runcard