Skip RNN TrackClassifier for bad tracks associated to the tau (Run4 dev)
Currently the RNN track classifier is running over all tracks associated to the taus, including also the tracks not passing the track selection requirement checked during the TauTrackFinder. For Run4, this should change and track classification should only be applied to tracks passing a minimum quality requirement -> this will save CPU avoiding running inference over tracks with bad quality.
This MR is inserting a flag to get this behavior and applying it when running with Run4 setup (see also slide 17 from https://indico.cern.ch/event/1613504/contributions/6807625/attachments/3180005/5657679/Slides_TauCP_20251125_ADM.pdf).
Edited by Antonio De Maria