updates from HLT users

[missirol]

This PR contains a few technical fixes and updates accumulated over time by people using this package for HLT JECs. As discussed with the JME conveners, the plan moving forward is to use cms-jet:master as reference branch for HLT (instead of branching out in forks as it happened in the past).

Attn: @sparedes @pallabidas @lathomas @kirschen

---

General comments:

• The updates in this PR should not interfere with any standard workflow used for Offline JECs.

• The changes are mostly the addition of options found useful for HLT studies, and minor technical fixes found along the way.

• A few changes go in the direction of removing HLT-specific choices inside the routines of JetMETAnalysis; I find it more transparent to add options for the user (for example, to decide which fit functions to use), rather than programs making choices under-the-curtain based on the name of a given jet collection. Arguably, JetMETAnalysis should not know the difference between HLT and Offline, it should just provide enough flexibility for a user to handle both.

• Some changes are just formatting ones (these were too many to undo in full, apologies); in general, it looks like it would be useful to add a couple of scripts to apply code-formatting rules to this package as a whole.

• Tested with CMSSW_11_2_0.

---

Main changes by file:

• addition of jet_synchfit_x__hlt: the standard program jet_synchfit_x was modified significantly over time, and understanding those changes requires more time. In order not to interfere with the standard tool (i.e. jet_synchfit_x), we would just add a separate program under a different name for the time being; the plan is to remove this in the near future, as soon as we figure out what are the relevant changes needed in jet_synchfit_x (if any).

• jet_response_analyzer_x.cc: added option to apply JECs on the fly. This addition seemed to be useful to apply the L1 JEC on-the-fly when preparing the inputs for the L2(+L3) fits, instead of creating a whole JRA NTuple with L1-corrected jets (which is what we were doing before, and at times requires a lot of disk space). Default behaviour of the program is unchanged.

• jet_synchtest_x.cc: added options to ignore NPV, and possibly replace it with NPU(in-time). These options are meant to circumvent certain choices that are customary for Offline, but might not always apply to HLT. The number of PVs at HLT is not as good a proxy for PU as it is for Offline, so it might be preferable to plot versus the number of (in time) PU interactions rather than N(reco-PVs). Default behaviour of the program is unchanged.

• JetResponseAnalyzer.cc: added option applyVtxCuts to the plugin. In the HLT case, one may want to ignore offline-like quality requirements in the count of NPV (those cuts aren't appropriate for HLT pixel vertices, for example). I think the default behaviour is preserved by the update in JetAnalyzers/python/Defaults_cff.py (ideally, one would have added a JetResponseAnalyzer::fillDescriptions method).

---

[missirol]

Kind ping for the review of this PR.

[missirol]

Closing this PR, since there was no review.