Split some IDPVM plots vs truthMu and actualMu

IDPVM used uniformly so far a pile-up variable defined as

puEvents = !m_truthPileUpEventName.key().empty() and truthPileupEventContainer.isValid() ?  static_cast<int>( truthPileupEventContainer->size() ) : pie.isValid() ? pie->actualInteractionsPerCrossing() : 0;

This is introducing in principle inconsistency between full pile-up truth and regular samples, as mu will be defined in the FPT samples as the exact number of simulated in-time pile-up interactions while in regular samples this will use the actualInteractionsPerCrossing from the EventInfo (number of in-time pile-up interactions estimated from lumi in principle).

This MR splits the plots of the number of tracks / vertices / efficiency vs truthMu and actualMu for clarity (and potential comparison in FPT samples).

Old unsupported ANT plots are also removed as they would otherwise require some interface update and since we don't use them any more...