Fix MultiTrackValidator efficiency summary and efficiency vs. deltaR plots

[makortel]

The recently (#9201) added efficiency summary plot (average efficiency for each input track collection) in MultiTrackValidator has a bug in putting also out-of-time TrackingParticles to the denominator. Also the computation of the DeltaR of minimum distance between any TP pair includes OOT TPs. This PR fixes both with the following developments (superseding #9956)

• Add new intimeOnly flag to TrackingParticleSelector (with value False set to everywhere to not to change current behaviour)
• Switch this flag to True for the TP selector used in MTV loop over TPs for efficiency plots to reject OOT TPs there
  • The efficiency denominators themselves are fine already, as all TP selectors used in MTVHistoProducerAlgoForTracker::fill_recoAssociated_simTrack_histos() already have signalOnly=True
  • This change alone would be enough to fix both of the issues with OOT TPs, but has a knock-on effect on histograms in Tracking/Track/<iter>/simulation which have also been filled from OOT TPs, and actually for those keeping more or less the current behaviour would be nice. Therefore,
• The histograms in Tracking/Track/<iter>/simulation are moved to Tracking/Track/simulation and filled only once per MTV instance (they are anyway the same for each iteration), and now from the set of all TPs without any selection (previously only those passing the TP selector in bullet 2 were used)
  • bunchxSIM is filled from both OOT and in-time TPs, while the others are filled only from IT TPs
• In addition, the same selection as for efficiency vs. DeltaR (same as efficiency vs. eta) is applied for efficiency summary plot, but now in two flavours: one with the pT cut (pT > 0.9 GeV) and other without pT cut.

Tested in 750pre6, expecting changes in effic_vs_coll, num_simul_coll, num_assoc(simToReco)_coll, and globalEfficiencies under Tracking/Track folder (num_*_coll also in other MTV instances), and effic_vs_dr, num_simul_dr, and num_assoc(simToReco)_dr in all MTV instances (in this case there is no change in the corresponding globalEfficiencies, as the integrals of numerator and denominator stay the same, only the DeltaR values change).

In noPU samples the efficiency summary plots should change like (black=old, red=new)

and in pileup samples like

In pileup samples the number of TPs and efficiency vs. DeltaR should change like (no changes expected in noPU samples)

@rovere @VinInn

[makortel]

I accidentally hit "submit" while still writing the description and running the final tests. All is fine now, so reopening.

[cmsbuild]

A new Pull Request was created by @makortel (Matti Kortelainen) for CMSSW_7_6_X.

Fix MultiTrackValidator efficiency summary and efficiency vs. deltaR plots

It involves the following packages:

CommonTools/RecoAlgos
PhysicsTools/RecoAlgos
SLHCUpgradeSimulations/Geometry
SimGeneral/MixingModule
SimGeneral/TrackingAnalysis
SimTracker/Common
SimTracker/VertexAssociation
Validation/RecoMuon
Validation/RecoTrack

@civanch, @Dr15Jones, @cvuosalo, @ianna, @mdhildreth, @monttj, @cmsbuild, @deguio, @slava77, @vadler, @danduggan can you please review it and eventually sign? Thanks.
@TaiSakuma, @schoef, @rappoccio, @Martin-Grunewald, @threus, @LBeck, @mmarionncern, @battibass, @ahinzmann, @jlagram, @jhgoh, @cnuttens, @prolay, @cerati, @trocino, @rociovilar, @abbiendi, @wmtan, @GiacomoSguazzoni, @rovere, @VinInn, @nhanvtran, @wmtford, @dgulhan, @jdolen, @gbenelli, @istaslis, @mariadalfonso this is something you requested to watch as well.
You can sign-off by replying to this message having '+1' in the first line of your reply.
You can reject by replying to this message having '-1' in the first line of your reply.
If you are a L2 or a release manager you can ask for tests by saying 'please test' in the first line of a comment.
@Degano you are the release manager for this.
You can merge this pull request by typing 'merge' in the first line of your comment.

[slava77]

@cmsbuild please test

[cmsbuild]

The tests are being triggered in jenkins.

[cvuosalo]

+1

For #10014 238a2c7

Bug fixes for MultiTrackValidator efficiency plots.

The code changes are satisfactory, and Jenkins tests against baseline CMSSW_7_6_X_2015-07-04-2300 display only the expected changes as shown in the PR description.