backport DNN cuts for mkFit 12_2_0_pre3

[JanFSchulte]

Backport of #36285 as part of the backport of mkFit from 12_1_0_pre3. Needs cms-data/RecoTracker-FinalTrackSelectors#11 to have the correspond DNN training file for the track selection.

[cmsbuild]

A new Pull Request was created by @JanFSchulte (Jan-Frederik Schulte) for CMSSW_12_1_X.

It involves the following packages:

• RecoTracker/IterativeTracking (reconstruction)

@jpata, @cmsbuild, @slava77 can you please review it and eventually sign? Thanks.
@felicepantaleo, @GiacomoSguazzoni, @JanFSchulte, @rovere, @VinInn, @ebrondol, @gpetruc, @mmusich, @mtosi, @dgulhan this is something you requested to watch as well.
@perrotta, @dpiparo, @qliphy you are the release manager for this.

cms-bot commands are listed here

• Backported from update working point for new DNN #36285

[tvami]

@cmsbuild , please test

[tvami]

backport of #36285

[cmsbuild]

+1

Summary: https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-324fe9/20918/summary.html
COMMIT: bbe8d33
CMSSW: CMSSW_12_1_X_2021-12-01-1100/slc7_amd64_gcc900
User test area: For local testing, you can use /cvmfs/cms-ci.cern.ch/week1/cms-sw/cmssw/36327/20918/install.sh to create a dev area with all the needed externals and cmssw changes.

Comparison Summary

Summary:

• No significant changes to the logs found
• Reco comparison results: 15748 differences found in the comparisons
• DQMHistoTests: Total files compared: 42
• DQMHistoTests: Total histograms compared: 2901440
• DQMHistoTests: Total failures: 26395
• DQMHistoTests: Total nulls: 6
• DQMHistoTests: Total successes: 2875017
• DQMHistoTests: Total skipped: 22
• DQMHistoTests: Total Missing objects: 0
• DQMHistoSizes: Histogram memory added: 0.027 KiB( 41 files compared)
• DQMHistoSizes: changed ( 136.874 ): 0.027 KiB JetMET/SUSYDQM
• Checked 177 log files, 37 edm output root files, 42 DQM output files
• TriggerResults: no differences found

[mmasciov]

@JanFSchulte, I believe that for a "full" backport one would also need a PR to cms-sw/cmsdist, to propagate the following change (update of tag) also to the IB/CMSSW_12_1_X/master branch:
https://github.com/cms-sw/cmsdist/blob/IB/CMSSW_12_2_X/master/data/cmsswdata.txt#L6
Otherwise, I think that the new cuts would be applied on top of the old training file.
In other words, https://github.com/cms-sw/cmsdist/blob/IB/CMSSW_12_1_X/master/data/cmsswdata.txt#L9) should be updated from RecoTracker-FinalTrackSelectors=V01-02-00 to RecoTracker-FinalTrackSelectors=V01-03-00?

[JanFSchulte]

Hi @mmasciov I think you are right and I have created the PR: cms-sw/cmsdist#7486.

[tvami]

@cmsbuild , please test with cms-sw/cmsdist#7486

[cmsbuild]

+1

Summary: https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-324fe9/20938/summary.html
COMMIT: bbe8d33
CMSSW: CMSSW_12_1_X_2021-12-02-1100/slc7_amd64_gcc900
User test area: For local testing, you can use /cvmfs/cms-ci.cern.ch/week1/cms-sw/cmssw/36327/20938/install.sh to create a dev area with all the needed externals and cmssw changes.

Comparison Summary

Summary:

• No significant changes to the logs found
• Reco comparison results: 18988 differences found in the comparisons
• DQMHistoTests: Total files compared: 42
• DQMHistoTests: Total histograms compared: 2901440
• DQMHistoTests: Total failures: 23012
• DQMHistoTests: Total nulls: 6
• DQMHistoTests: Total successes: 2878400
• DQMHistoTests: Total skipped: 22
• DQMHistoTests: Total Missing objects: 0
• DQMHistoSizes: Histogram memory added: 0.035 KiB( 41 files compared)
• DQMHistoSizes: changed ( 136.793 ): 0.023 KiB JetMET/SUSYDQM
• DQMHistoSizes: changed ( 136.874 ): 0.012 KiB JetMET/SUSYDQM
• Checked 177 log files, 37 edm output root files, 42 DQM output files
• TriggerResults: no differences found

[slava77]

+reconstruction

for #36327 bbe8d33

• code changes are (essentially) the same as the master branch variant update working point for new DNN #36285 (which was already merged)
• jenkins tests pass and comparisons with the baseline show differences in phase-1 workflows (2017 through 2023; as expected). A spot check of a few distributions in wf 136.874, the counts of tracks and different kinds of PF objects seems to be very similar (tracks are identical; PF varied due to additional updates in 12_2_X) https://cmssdt.cern.ch/SDT/jenkins-artifacts/baseLineComparisons/CMSSW_12_1_X_2021-12-02-1100+324fe9/47209/validateJR/all_OldVSNew_RunEGamma2018Cwf136p874/ vs https://cmssdt.cern.ch/SDT/jenkins-artifacts/baseLineComparisons/CMSSW_12_2_X_2021-11-29-2300+5790f3/47148/validateJR/all_OldVSNew_RunEGamma2018Cwf136p874/

[cmsbuild]

This pull request is fully signed and it will be integrated in one of the next CMSSW_12_1_X IBs (tests are also fine) and once validation in the development release cycle CMSSW_12_2_X is complete. This pull request will now be reviewed by the release team before it's merged. @perrotta, @dpiparo, @qliphy (and backports should be raised in the release meeting by the corresponding L2)

[slava77]

this PR is a catch-up to #36315, also proposed for the backport to 12_1_X
and, IIUC, the last reco code change needed to be backported to 12_1_X at this point.

[tvami]

urgent

• needed for the the 12_1_X

[perrotta]

+1

• Master version succesfully tested
• It changes reco, but there are no 12_1_X productions with reco started yet
• It goes together with Update RecoTracker-FinalTrackSelectors tag (backport of #36285) cmsdist#7486