add the local b-field to the TrackState and store the correct curvature for that position

[bloodyyugo]

Currently, the curvature at the different kalman track states is corrected such that it gives the correct pT if you use the B-field at the center of the magnet (~0.85T for 2021/3.7GeV run). So, the curvature stored at the target (where the field is ~0.7T) is currently incorrect.

The reason we do this is because in the LCIO code the Track/TrackState just saves the track parameters & reference position but no momentum or local bfield to calculate the momentum.

This issue, along with corresponding changes to hps-lcio and lcsim-hps, adds the local bField (at the reference) to the TrackState object so that it is saved and readout by lcsim-hps (for hps-java recon/reading, JeffersonLab/hps-lcsim#1) and hps-lcio (for reading hps-java output and writing hpstr tuples, JeffersonLab/hps-lcio#13).

[bloodyyugo]

This was completed with #1101