Update Cabibbo-favoured charm lines based on early '24 data

This MR includes some additional cuts on Cabibbo-favoured charm lines to reduce their rate, effectively without affecting signal. It's based on the data tuples that we had from early '24, and was partly presented in https://indico.cern.ch/event/1427005/contributions/6020304/subcontributions/490487/attachments/2885072/5056274/240626_D2Ksh.pdf

Expect that this cuts the rate of D -> Ksh in half, tested on 90k data events. Meanwhile, removed the BPVIPCHI2 cut from the d_to_hhh combiner...