Notable changes:
* In general, reda should run better on newer python/package versions
* Add a type3-pseudosection plotting routine: This routine computes data
pseudo locations based on the sensitivity-based center of mass, based on
modeling results from CRMod.
* Add .compute_K_numerical to the Base Container
* [seit container] add no_norrec=False parameter to .import_crtomo
* [base container] add reassess_norrec=True parameter to .filter
reassess_norrec : bool (True)
if True, then recompute normal-reciprocal differences after
applying the filter.
* [base container] add .get_norrec_pairs(self, test_column='rdiff')