Merge feature/kalman, TMS_Bar and TMS_Kalman small changes

[LiamOS]

config/TMS_Default_Config.toml:

• FirstCluster now set to 'false' by default

TMS_Bar:

• Added GetBarWidth() and GetBarLength(), Width/Length are always the short/long bar dimension respectively

TMS_Kalman:

• cout and cerr output much more tolerable
• Check on initial Y slope for some tracks
• Better check on values while constructing nodes
• Can now handle horizontal (X) layers in theory, untested
• Detector noise matrix calculated using layer geometry (from TMS_Bar additions above)
• Chi2 calculated per-node, can be summed for a track
• KalmanErrorDetVol variable added, 1 if Kalman filter predicts a node outside the scintillator, 0 otherwise

TMS_Reco

• SetStart/EndDirection is previously only called for tracks when Kalman was run, now set it as start-end (old way) if Kalman isn't run.

TMS_Track

• SetStartDirection/SetEndDirection now automagically normalise the vectors given to them.
• Chi2 variable added per track, to be filled by Kalman (in future maybe chi2_mu / chi2_amu?)
• KalmanErrorDetVol propagated to TMS_Track object

TMS_TreeWriter

• RecoTrackHitPos and related now include time, and thus have an extra index ([3] -> [4]).
• Added Chi2 to output files, filled by RecoTrackChi2 in TreeWriter.cpp
• KalmanErrorDetVol included in output trees per track

[LiamOS]

Should be behaving substantially better with this, but not perfect. Still leaves the detector sometimes. Want to wait for some feedback for Asa before tackling this fully.

[LiamOS]

Validation plots from one file located in /exp/dune/data/users/losulliv/dune-tms/Validation/Reco_Tracks/MicroProdN1p1_NDLAr_1E18_RHC.spill.nu.0000120.EDEPSIM_SPILLS_TMS_RecoCandidates_Hough_Cluster1_images/

For future PRs, should we define some minimal set of required validation plots that should be presented, or do it depending on the content of the PRs?

[AsaNehm]

Solved merging conflict by removing the sanity check in TMS_Bar that was still in in this branch but removed on main