lordome
commented
2 years ago Vertex identification as of now is just okay - the traces are quite well identified, but the approach taken is much different than the one taken before
The 'beamTracks' are no more identified for their energy and length, as some information was lost in 'scattering' events (the vertex could not be reconstructed if the 'beam' track was not fitted, as in the event would be present a single track). Now the identification principle is based on the fact that some tracks have a minimum / maximum Z that is not inside the 'beam' region. If some pileup is present, the information on the height can be exploited to classify such traces as 'not fittable'.
Some fine tuning might me needed a posteriori to address whether the number of traces included in the vertex is really the number of traces in the event. Indeed, maybe the trace associated to the beam particle 'entering' in the vertex might have a continuation after the vertex, or can just stop there as the ejectiles might have all large emission angles. In the first case, the beam-like trace has to be counted inside the number of ejectiles, while in the latter case it has to be discarded. For example, a simple check on whether points belonging to the trace before and after the vertex are present can be included. In any case, such fine tuning is typical of the single experiment, and therefore is not of interest for the current status of the work.
There are a high number of events with a null vertex (non physical) (-100)
Check which traces are included into one vertex