[Halld-physics] Kinematic Fitter Overhaul

[Paul Mattione]

I just overhauled the kinematic fitter to fix a bunch of long-standing issues: 

1) Decaying particles and constraints now handled in a more intuitive manner. 
2) P4 fit no longer required to do vertex fits with decaying particles. 
3) No more issues with some constraints not working (e.g. pi0 mass constraints OK now). 
4) (All?) Corner cases now work. 
5) Improved interface. Code cleaned up significantly.

This has been tested for the following channels: p2pi, p3pi, pomega, b1pi, klambda, lambda, xi-
This has been tested with the following plugins: b1pi_hists, p2pi_hists, p3pi_hists, kinfit_test (my home folder)

If you're not using the fitter directly, the only change you need to make to your DReaction is to replace:

Set_ApplyKinFitMassConstraintOnInitialParticleFlag() -> Set_KinFitConstrainInitMassFlag()

If you are using the fitter directly, the interface has changed, so you'll have to rewrite your code. See the documentation link for instructions.

Documentation is at: https://halldweb.jlab.org/wiki/index.php/Mattione_GlueX_Kinematic_Fitting

Also, the following histograms were added to the main library:
Pulls vs p/theta/phi
Confidence level vs p/theta/phi
Direct invariant mass of several particles in a step (e.g. rho)
Lifetime of decaying particles in their rest frame (if vertex fit performed)


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.jlab.org/pipermail/halld-physics/attachments/20160209/3d9b46ae/attachment.html>