Fix couting of pre- vs post-finalize objects in AH
In the merging logic of the AH, there is a check that the total number of AOs in an input YODA file is twice the number of raw AOs, else the file is declared corrupt and skipped in the merging.
This does not account for TMP/
histos which are only kept at pre-finalize level and not included as post-finalize version (they are intended to be temporary after all).
This change set fixes the self-consistency check and also replaced regex_match
with regex_search
to pick up substring matches.