Open
Description
Things to be checked/taken care of, especially with respect to the 102X 2016/7/8 differences. I've grouped them by object, since often they are common across years.
Please comment/edit with any missing items!
Some things are only important post-ntupleisation, which have been marked with [analysis]. Everything else is assumed to be important for making ntuples.
PdmV recipe: https://twiki.cern.ch/twiki/bin/view/CMS/PdmVRun2LegacyAnalysisSummaryTable
GENERAL
Datasets/PPD recommendations UL20 MiniAODv2 MiniAODv1
- propagation of era flag
- preferred solution: resolve all
import *and move into generator function --> @hjabusch [106X (& 102X)] reduce ntuple generator imports to needed functions only #1593 - alternatively: add explicit statement (encapsulate?) at beginning of each ntuplewriter file
- preferred solution: resolve all
- update CI tests to Summer20 samples --> @mschrode ([106X] Update CI test files to UL for 2017 #1590 [106X] Update CI test files for UL #1595)
- update crab production scripts --> @MatthiesC [106X] better request names and more commenting in crab_template.py #1587 More changes to crab_template.py #1591
- storage path
- PUPPI version v15 --> adjust code to not re-cluster with v15 but use directly from MiniAODv2 samples --> @kdeleo, @stalbrec, @abenecke [106X] Removing UpdatePuppiTuneV15 from ntuple_generator #1592
- Update to latest --> @jabuschh
- CMSSW_10_6_28 HN [106X] update CMSSW version from 10_6_26 to 10_6_28 #1589
- Global Tags & Eras [106X] update global tags for eras #1585
- prefiring weights for UL Twiki recipe --> [106X] prefiring weights for UL17 in ntuple generator #1568 @MatthiesC
- Updated PU reweighting (? )
- not relevant for the production, PU weights not stored in ntuples
- HN
- Change of SF to JSON format
- needs (possible) future adaption of code --> probably not ready by all POGs
- not relevant for production, concerns running on ntuples
- Update to higher CMSSW version (needed for egamma reco tools, Puppi v15) --> done in [106X] Update to CMSSW_10_6_26 #1572
- Review list of variables included in the ntuples (based on the histograms being compared during CI): many not filled/not meaningful values @mschrode for MC: empty, zero RMS, zero value
- Why is geninfo.systweights not filled? @stalbrec problem of an old sample
- Drop support for <= 102X datasets? --> keep for the time being (currently 102X ntuples still used in analysis development, for analyses close to final)
JETS
- Fix partonFlavour in jetsAk4Puppi (broken when updatePuppiv14 introduced in [106X] updated to v14 with CMSSW 10_6_13 #1519 ): related to reclustering at ntupling stage, fixed in PUPPI v15
- either miniAOD v2, or
- update to v15 on the fly
- AK8CHS/PUPPI PUJetID not stored (jet collection and top collection) --> @apaasch, not defined
- Update list of btaggers - surely some outdated stuff in there? @finnlabe @jabuschh [106X] remove b-tagger combined secondary vertex (CSV) #1581
- Add new taggers (eg. ParticleNet) @stalbrec
- ParticleNet added for AK8 jets in [106X] Updating to recommended CMSSW version and adding ParticleNet #1565
- Add mass regression --> @stalbrec
- GenJets
-
genjetsAk8Substructureempty or same value: subjet info empty → ungroomed AK8 GenJets → no subjets → use genjetsAk8SubstructureSoftDrop collection instead -
slimmedGenJetsAK4/AK8.genparticles_indices empty --> [switches](https://github.com/UHH2/UHH2/blob/RunII_106X_v1/core/python/ntuple_generator.py#L2506) set tofalse` as default- m_genparticles_indices holds the indices (integers) of all genparticles related to the genjet
-
slimmedGenJetsAk8(works for Ak4) pdgId, hadron/partonFlavour same value --> corresponding jet-parton matching not done (this condition alwaysfalse) --> @jabuschh: concerns all AK8 GenJet collections: slimmedGenJetsAK8, genjetsAk8Substructure, genjetsAk8SubstructureSoftDrop - pdgid at default value: expected behavior
- partonFlavour, hadronFlavour at default value → genjet flavor matching is not working (&gen_jet == aJet always false for AK8 GenJets, link), is does work for AK4 → still not really clear unfortunately
-
slimmedGenJetsAk8subjet info not available --> @jabuschh: ungroomed AK8 GenJets → no subjets → use genjetsAk8SubstructureSoftDrop collection instead (same as 1.)
-
- JME IDs/recommendation --> will be done later, not relevant now @anmalara
- JetID UL recommendations
- JEC/JER recommendations
- Drop btag_info stuff? Not filled / used - potentially adds up a lot (~35 * 0.0038 kB / jet collection, * ~8 collections = 1 kB /event, relative to ~ 9 kB / event MC) --> done with [106X] Remove JetBTagInfo class #1544
- Fix L1Jet_seeds being empty e.g. validation plot @finnlabe --> expected behaviour, not filled for MC, being updated with imminent PR related to L1 info/objects
LEPTONS
- Lepton_keys missing for HOTVR / XCONE Lepton_keys missing in 102X for HOTVR / XCONE #1468 --> @MatthiesC
- missing lepton keys: jet-lepton cleaning can only be done via DeltaR-based cleaning OK for relevant analyses
- Add flag if lepton can be matched to L1 lepton --> @finnlabe @jabuschh [106X] hotfix for L1-reco matching (PR#1579) #1580
- Use EGamma PostRecoTools [106X] Replacing our Egamma code by POG postrecotools #1571 --> @finnlabe
- drop support for producing 102X ntuples 2016 and 2017
- check that we can still run on EOY ntuples all years
- EGammaID: use consistently tags in ntuples, deprecate on-the-fly ID classes
- Add the latest versions (Tags, IDs, calibrations etc.) for muon/ele and similar recommendations --> @MatthiesC (details see comment below)
- See https://hypernews.cern.ch/HyperNews/CMS/get/muon/1747.html --> @MatthiesC only related to JSON format SF files
MET
- Update list of METfilters --> at least one,
BadPFMuonDzfilter only in Summer20 samples --> @jabuschh [106X] update MET filters for UL datasets #1584 -
slimmedMETsfilled with 0 --> @abenecke
EGAMMA
-
slimmedPhotonsUSERzero [106X] added flag for photon sources and switched them on (as default) #1576 --> @jabuschhm_ptError = 0, m_etaError = 0, m_phiError = 0--> algo only valid for electronsm_patParticleIsocommented out, deprecated --> default value
GENPARTICLES
- several empty/zero variables --> @stalbrec
GenParticles.spinwe don't store it (LHE event product not available for every sample)genInfo.pdf_xPDF1/2filled (0 in miniAOD)
-
GenParticles.hadron(parton)Flavourzero RMS/same value --> @jabuschhGenParticleinherits fromFlavourParticle, which defines fields hadron/partonFlavour
Metadata
Metadata
Assignees
Labels
No labels