Empty grid for DeepTauId

[jfernan2]

While implementing basic guards to solve the empty input problem in DeepTauId there were still empty grids which need to be investigated with Tau experts.

#44333 (comment)_

[jfernan2]

assign reconstruction

[cmsbuild]

New categories assigned: reconstruction

@jfernan2,@mandrenguyen you have been requested to review this Pull request/Issue and eventually sign? Thanks

[cmsbuild]

cms-bot internal usage

[cmsbuild]

A new Issue was created by @jfernan2.

@makortel, @antoniovilela, @rappoccio, @Dr15Jones, @smuzaffar, @sextonkennedy can you please review it and eventually sign/assign? Thanks.

cms-bot commands are listed here

[jfernan2]

type tau

[jfernan2]

@cms-sw/tau-pog-l2 FYI

[jfernan2]

@valsdav could you please post here instructions to reproduce the issue in case they had changed from original issue ?
Thank you

[valsdav]

Hi! Here is the reproducer to get the event with the empty grid.
From #44333 (comment)

• cmsrel CMSSW_14_0_0; cd CMSSW_14_0_0/src; cmsenv
• Pick event:

edmCopyPickMerge outputFile=pickevents.root eventsToProcess=367131:196942831 inputFiles=/store/data/Run2023C/DisplacedJet/RAW/v1/000/367/131/00000/9f3f571f-6dc9-4bda-a68b-5d1b9a5fc3ac.root

• Run:

cmsDriver.py step2 --conditions auto:run3_hlt_relval --data --datatier FEVTDEBUGHLT --era Run3_2023 --eventcontent FEVTDEBUGHLT --filein file:pickevents.root --fileout file:step2.root --nStreams 4 --nThreads 8 --number -1 --process reHLT --python_filename step_2_cfg.py --step L1REPACK:Full,HLT:@relval2024 --accelerators cpu

• setup deetau log level to 2

You will see empty grids in the input preparation.
If you merge on top my PR #44455 the inference will be skipped on the empty inputs.

[jfernan2]

ping @cms-sw/tau-pog-l2

[jfernan2]

ping @cms-sw/tau-pog-l2 @mbluj @Ksavva1021

[mbluj]

@kandrosov @brallmond could you also take a look, please?

[kandrosov]

@mbluj I have identified the tau that has the empty signal grid in the example event pointed out above:

tau pt=52.9444 eta=-0.0898154 phi=-0.68437
innerSignalConeRadius = 0.0566632
signalTauChargedHadronCandidates
  pt = 32.8809, eta = -0.177717, phi = -0.669344, mass = 0.13957, deltaR = 0.0892
    charge = 0 (pdgId = 130)
    charged PFCandidate (3:2826:1039): pt = 32.8809, eta = -0.177717, phi = -0.669344 (pdgId = 130)
    reco::Track: N/A
    neutral PFCandidates: N/A
    position@ECAL entrance: x = 163.066, y = -129.112, z = -44.1041 (eta = -0.21049, phi = -0.669706)
    algo = PFNeutralHadron
signalPiZeroCandidates
  pt = 11.7301, eta = 0.0901726, phi = -0.745689
    daughter #0 (PFGamma): pt = 3.92827, eta = 0.0625999, phi = -0.757536, deltaR = 0.169
    daughter #1 (PFGamma): pt = 2.98864, eta = 0.0810623, phi = -0.714655, deltaR = 0.174
    daughter #2 (PFGamma): pt = 3.44201, eta = 0.116514, phi = -0.774948, deltaR = 0.225
    daughter #3 (PFGamma): pt = 1.37547, eta = 0.122283, phi = -0.70606, deltaR = 0.213
  pt = 7.9762, eta = 0.0110623, phi = -0.656186
    daughter #0 (PFGamma): pt = 7.9762, eta = 0.0110623, phi = -0.656186, deltaR = 0.105

Indeed all signal PF candidates have deltaR > innerSignalConeRadius, where innerSignalConeRadius is computed with getInnerSignalConeRadius.
This happens for HLT path HLT_VBF_DiPFJet45_Mjj550_MediumDeepTauPFTauHPS45_L2NN_eta2p1_v2, which uses as an input hltHpsPFTauProducer -> hltHpsPFTauProducerSansRefs -> hltHpsCombinatoricRecoTaus.
In hltHpsCombinatoricRecoTaus signalConeSize = cms.string('max(min(0.1, 3.0/pt()), 0.05)'). So I don't understand how it is possible.