hadronic recon methods performance

[c-dilks]

Seems like the resolutions from hadronic reconstruction methods are wider than expected

• see https://github.com/c-dilks/largex-eic/pull/112 for attempts and resolving this, and a summary of the current understanding of the issue
• the determination of the hadronic final state may be a culprit:
  • the invariant mass of the hadronic final state from Kinematics::GetHadronicFinalState tends to be smaller than that from Kinematics::GetHadronicFinalStateTrue, especially at low x.
  • requiring PID to be smeared in GetHadronicFinalState (i.e., rejecting events which were not PID-smeared), causes resolutions to be much worse.

[c-dilks]

close for now, since ultimately we will need to think more carefully about how to handle the hadronic final state