D0->hh filtered MC (prompt) 2024, bestPV and allPV DTF

Analysis production for filtered MC 2024 D0->hh samples (both prompt charm and from-B EvtTypes).

Based on !3398 (merged), and supersedes it.

The difference from them is that now DTF with PV constraint is run with fit_all_pvs once equal to True and once equal to False. In the previous version, only False was used. We would like to have both to study possible misbehaviours of DTF we see in a tiny fraction of events.

Only tagged samples are tupled, since MC is filtered on prompt-tagged lines.

Section related to variables obtained from MCReconstructed is currently commented since it always gives *_RECOSTRUCTED = -1. To be fixed in a future AP (maybe, it is not really needed).

Only prompt charm EvtTypes are processed, for the moment.