CKF assumes one track per seed

[osbornjd]

Currently the CKF takes the output and puts the information directly into an Acts::MultiTrajectory type object, which an additional algorithm ParticlesFromTrackFit reads from and translates into an edm4eic object. As discussed in the track reconstruction meeting today, the interpretation currently assumes that the CKF only finds one branch per input track seed as shown here and here. Currently we are basically taking the longest track branch per track seed as "the track", which is probably a very good assumption in no background environments. This likely does not matter for tracking without background as there are probably very few cases where multiple branches per input track seed are found. However, this will (my suspicion is) be a very bad assumption once background is implemented. We should consider modifying this to including all branches for a given track seed, which may then require some algorithm to identify which branch is the best and should be taken as "the track fit" associated to a given track seed.

[osbornjd]

Related to this, what is this piece of code doing? It appears to be taking all of the found branches from the CKF and then iterating over them as if they are track states, then setting the reconstructed particle momentum and charge to be that of some track state? Unless I am misunderstanding, this is mixing track branches and states which may lead to undetermined behavior...