[Halld-offline] Updates to the Analysis Library

[Paul Mattione]

I've made a few updates to the analysis library, which are listed below.  The first one will break your plugins, but it is quite easy to fix:

1) I changed the names of the methods of a few of the pre-combo cuts you can make in your DReaction.  If you update your version of the analysis library, just do the following replacements and your code will work fine (also checkout the latest version of MakeReactionPlugin!!):
  a) Set_MinIndividualTrackingFOM   ->   Set_MinTrackingFOM
  b) Set_MinIndividualPIDFOM   ->    Set_MinChargedPIDFOM
  c) Set_MinIndividualNeutralPIDFOM    ->   Set_MinPhotonPIDFOM

2) Massive neutral particle (e.g. Neutron, KLong) reconstruction is now supported, although kinematic fitting is not (yet).  For photons, the shower energy is used as the photon momentum, which is used with the time measurement to calculate beta a the PID FOM.  For massive neutral particles, the shower energy has a complex relationship to the particle energy and is thus not used.  Instead, the timing is used to calculate beta which is used to set the momentum, and the PID FOM is undefined.  Thus cuts on the neutron PID FOM are ignored.  The safest way to select non-photons is to do a kinematic fit and let the mass & p4 constraints eliminate most of the photon background.  Eventually, we can feed information about shower depth into the BDT as well.  Note that at the moment, the neutron momentum resolution (which is directly tied to the calorimeter timing resolution) is very broad and off-center.  

3) A new reaction-independent histogram action, DHistogramAction_ReconnedThrownKinematics, has been defined and included in the monitoring_hists plugin.  It histograms the kinematics of the thrown particles that are matched (with a modifiable matching FOM cut) to reconstructed particles.  This is useful for quickly examining the MC acceptance.  

4) To help clean up the junk tracks from the histograms, there was a cut in DHistogramAction_DetectedParticleKinematics requiring that the reconstructed tracks match to thrown tracks.  This cut has now been removed.  Therefore, all counts in these histograms are inflated, but now more accurately represent what we would think we have in experimental data.  In this action, there are still built-in (modifiable) cuts on PID and tracking FOM (each +/- 5 sigma).

 - Paul