Reverting to AK4 CHS jets for DQM PNet monitoring

[marinakolosova]

This is a fix to PR 416555

• After discussing with @missirol, the ParticleNet monitoring should only be in HLTMonitor sequences.
• After discussing with @rgerosa and Stephane Cooperstein, we decided to revert to using CHS jets. The PNet training used currently in the DQM (older training) is not supported for Puppi, only for CHS jets. The new training for 13-0-X is only evaluated at the MiniAOD level, not at RECO, and for this reason, we revert to using CHS jets for the DQM monitoring.

Backports needed in:
CMSSW_13_1_X
CMSSW_13_0_X

[cmsbuild]

+code-checks

Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-41694/35566

• This PR adds an extra 16KB to repository

[cmsbuild]

A new Pull Request was created by @marinakolosova (Marina Kolosova) for master.

It involves the following packages:

• DQMOffline/Trigger (dqm)

@nothingface0, @emanueleusai, @cmsbuild, @pmandrik, @syuvivida, @tjavaid, @micsucmed, @rvenditti can you please review it and eventually sign? Thanks.
@mtosi, @Fedespring, @missirol, @HuguesBrun, @jhgoh, @trocino, @cericeci, @rociovilar this is something you requested to watch as well.
@perrotta, @dpiparo, @rappoccio you are the release manager for this.

cms-bot commands are listed here

[perrotta]

please test workflow 140.031

[perrotta]

Thank you @marinakolosova !

[perrotta]

type bug-fix

[missirol]

Thanks @marinakolosova !

[scooperstein]

Hi,

Just to add some further explanations here - the offline ParticleNet jet tagging for AK4 jets was updated recently (including regression+tau ID/reconstruction, on PUPPI jets) and backported to CMSSW_13_0_X #40745. The new network runs on miniAOD updates and cannot be run in the RECO step, whereas the previous version could be evaluated on either RECO or miniAOD inputs, and was trained on CHS jets. Since the DQM workflow is from the RECO/AOD as far as I understood, the new tagger is difficult to integrate. As a solution, we use the older offline tagger on CHS jets for the DQM - the b-tagging outputs should be highly correlated.

Please hold off on integrating this PR until @marinakolosova has a chance to fully test this.

[cmsbuild]

+1

Summary: https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-614286/32668/summary.html
COMMIT: 4785102
CMSSW: CMSSW_13_2_X_2023-05-16-1100/el8_amd64_gcc11
User test area: For local testing, you can use /cvmfs/cms-ci.cern.ch/week1/cms-sw/cmssw/41694/32668/install.sh to create a dev area with all the needed externals and cmssw changes.

Comparison Summary

Summary:

• You potentially added 97 lines to the logs
• ROOTFileChecks: Some differences in event products or their sizes found
• Reco comparison results: 12 differences found in the comparisons
• DQMHistoTests: Total files compared: 48
• DQMHistoTests: Total histograms compared: 3222710
• DQMHistoTests: Total failures: 9
• DQMHistoTests: Total nulls: 0
• DQMHistoTests: Total successes: 3222679
• DQMHistoTests: Total skipped: 22
• DQMHistoTests: Total Missing objects: 0
• DQMHistoSizes: Histogram memory added: -40.097 KiB( 47 files compared)
• DQMHistoSizes: changed ( 139.001 ): -40.097 KiB HLT/HIG
• Checked 210 log files, 162 edm output root files, 48 DQM output files
• TriggerResults: no differences found

[emanueleusai]

+1

[cmsbuild]

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

[marinakolosova]

Hi all, from our side everything looks ok, the PNet scores for AK4 CHS are being evaluated and the workflow is running as expected. No further update is needed for this PR, I will proceed with the backports. Thank you.

Hi,

Just to add some further explanations here - the offline ParticleNet jet tagging for AK4 jets was updated recently (including regression+tau ID/reconstruction, on PUPPI jets) and backported to CMSSW_13_0_X #40745. The new network runs on miniAOD updates and cannot be run in the RECO step, whereas the previous version could be evaluated on either RECO or miniAOD inputs, and was trained on CHS jets. Since the DQM workflow is from the RECO/AOD as far as I understood, the new tagger is difficult to integrate. As a solution, we use the older offline tagger on CHS jets for the DQM - the b-tagging outputs should be highly correlated.

Please hold off on integrating this PR until @marinakolosova has a chance to fully test this.

[perrotta]

+1