https://github.com/cms-sw/cmssw/pull/17308

Datasets

When there were multiple AOD datasets for a given era, a version label has been appended to the 03Feb2017 to distinguish between the outputs coming from the different inputs (and which should not be confused with the version of the MINIAOD dataset, the final -v(number) in the name, which is just the number of attempts it took to get the dataset successfully processed by the production system).

• For Run2016B, you will find 2 datasets called Run2016B-03Feb2017_ver1 and Run2016B-03Feb2017_ver2 ; these map to the RAW Run2016B-v1 and Run2016B-v2
  You can normally skip the ver1 dataset as it does not contain any runs in the golden JSON.
• For Run2016H, you will find two datasets called Run2016H-03Feb2017_ver2 and Run2016H-03Feb2017_ver3; these map to the PromptReco-v2 and PromptReco-v3 AODs.
  You should use both datasets, as they do not overlap.

EGM

READ THIS: It would be extremely advisable to store the bool particleFlowEGammaGSFixed:dupECALClusters and if ecalMultiAndGSGlobalRecHitEB:hitsNotReplaced is empty or not in any analysis job. If particleFlowEGammaGSFixed:dupECALClusters = true or ecalMultiAndGSGlobalRecHitEB:hitsNotReplaced is not empty, the fix may not have worked for that event and it needs to be manually recovered or at least scrutinized further. This events should be extremely rare but are possible. If you want to save which objects had a gain switch in at least one of its hits, see the userInts in this section.

The standard collections store the objects after the ECAL slew rate mitigation . We also store the old collections for comparison. A technical description of the changes can be found in a dedicated wiki.

Short Description of E/gamma Collections

This only details the fixed collections. To get the original collections add "BeforeGSFix" to the producer name as shown above.

Key Caveats with the Fix

This fix is mostly but not completely transparent to the user. Some small things changed. Read this section for detailed information on all the caveats associated with the fix.

Expert Information

show expert information... Hide description...

electron.userInt("hasGainSwitchFlag") == 1;
photon.userInt("hasGainSwitchFlag") == 1;

electron.superCluster()->seed()->seed().rawId() == oldelectron.superCluster()->seed()->seed().rawId();

Muons and PF Candidates

In this re-miniAOD, the filters for bad and duplicate muons advertised on physics-validation/2786 are used.

Event flags

Three flags are saved in the event:

• Flag_badMuons: the event contained at least one PF muon of p_T > 20 GeV that is flagged as bad
• Flag_duplicateMuons: the event contained at least one PF muon of p_T > 20 GeV that is flagged as duplicate
• Flag_noBadMuons: the event does not contain any PF muon of p_T > 20 GeV flagged as bad or duplicate (i.e. the event is safe)

Muons

The slimmedMuons collection contains all muons, as before.

However, the muons flagged as bad or duplicate are now no longer marked as PF muons; thus, isPFMuon() will return false on them (and consequently also isLooseMuon, isTightMuon, etc.). Nothing else of the muon is changed (e.g. the 4-vector is as given by the particle flow)

You can access the old value of the PF id flag via muon.userInt("muonsCleaned:oldPF") . If you want to make the muon PF again, you can copy it by value and then do muon.setType(muon.type() | reco::Muon::PFMuon).

ParticleFlow candidates & PUPPI weights

The bad and duplicate muons are removed from the packedPFCandidates list, and put in a separate collection packedPFCandidatesDiscarded; thus, the sourceCandidatePtr for bad or duplicate slimmedMuons= will therefore point to that second collection.

The PUPPI weights are computed from the cleaned PF candidates. The discarded PF candidates do not have a valid PUPPI weight.

MET Recipes

The color scheme of the instructions is as follows Hide

General Comments for MET Recipes

We strongly encourage everyone to switch to the CMSSW_8_0_26_patch1 release, which has been used for re-miniaod production. After switching to this release you will merge the default MET recipe.

   git cms-merge-topic cms-met:METRecipe_8020 -u

In this MET recipe you will find:

• Latest / greatest MET Filters
• Bug fixes related to MET Significance computation
• Implementation of the Smeared MET and JER uncertainties
• Bug fixes for puppi met specifically on the puppi photon treatment

You can find the full PR and the details here. There is no need to merge anymore any further recipes individually. However the instructions on how to use the different filters & functions specified on twiki1 and twiki2 will still hold.

What is in re-miniaod Data?

In the re-miniaod data you will find 4 different pf MET collections. This is to ensure the user has the maximum flexibility to use any level of e/gamma or muon corrected MET. The full list and the description is provide below. Please note that, for the e/gamma corrected MET, the global event description is not possible (i.e. jets are not e/gamma corrected). This is because e/gamma corrections are not performed at the pf candidate level. Therefore, e/gamma corrected MET can be viewed as "Type1 e/gamma corrected MET".

For the puppi MET, the only collection saved is the fully corrected MET. If user would like to "un-correct" puppi MET, they can use the following variables:

• puppiMETEGCor "corX" , "corY" and "corSumEt"
• puppiMETMuCor "corX" , "corY" and "corSumEt"

MET Q&A:

• Q: I am looking at the re-miniaod data, what kind of MET should I be using?
• A: First, please note that we here propagate corrections provided by MUO and EGM POG. The most correct MET collection is slimmedMETsMuEGClean, this is the collection corrected by both e/gamma and muon effects. But what if you are seeing some tails or inconsistencies? It could be that:

• (1) Bad Muon Filters on the pf candidates have removed a good pf muon candidate -> new tails
• (2) Bad Muon Filters on the pf candidates didn’t catch all the bad pf muon candidates -> not fully cleaned tails
• (3) EGamma GS Fix wasn’t optimal, some sort of mis-match has happened -> new tails.

If case (1) has happened, you can go back to slimmedMETsEGClean + use the Bad Muon Filters to reject the events that were tagged (you loose efficiency for the mis-tagging part but you don’t end up with tails)

If case (2) has happened, you should report the events. That means we need better filters for the legacy re-reco.

If case (3) has happened but the muon fix was behaving well and your analysis is dependent on high pt electrons and photons, you should re-run the recipe given on this twiki about the e/gamma corrected MET. We strongly encourage all analyzers to implement this recipe if they are sensitive to the gain switch issue.

• Q: Is it OK for me to use out of the box MET in data?
• A: If you are only interested in the type1 corrected MET and not sensitive to the gain switch issue, then yes! In data, you can use out of the box MET. However, if you are interested in using the newly implemented JER uncertainties on MET, or updated MET Significance, or you are sensitive to the e/gamma gain switch issue, we recommend you follow the recipes given in this twiki.

• Q: What about MC? What do I do there?
• A: We did not have the e/gamma gain switch problem in the monte carlo, therefore any fix / recipe for this situation is not relevant. However, (although in much much smaller magnitude) one can have “bad muons” also in the simulation. If you want to check the effect of the filters or bad muon corrections on MET on your simulation, you can follow the recipe given under “how do I correct mc on the fly for bad muons”. This is not mandatory but is advised. You can think of this suggestion as our regular suggestions of checking the filter efficiencies on MC.

• Q: When should I JEC correct and when should I re-cluster MET?
• A: Are you doing something special with the pf candidates? If not, then re-correcting the met based on JECs is enough for you. The re-clustering jets/mets option in the recipe is for expert use, or if you want to change your pf candidate collection with some analysis specific candidates. For most analyzers re-correcting the MET for the latests JECs is enough.

How to re-correct MET based on bad muons on the fly for MC (Release: CMSSW_8_0_X, X>=26_patch1)

In this recipe, we are creating a new muon collection based on the tagged bad muons, and cleaning the pf candidates with respect to these bad muons. Using the cleaned pf candidates we will be re-clustering MET.

Full recipe is below Hide the recipe

   ## Following lines are for default MET for Type1 corrections.
   from PhysicsTools.PatUtils.tools.runMETCorrectionsAndUncertainties import runMetCorAndUncFromMiniAOD

   # If you only want to re-correct for JEC and get the proper uncertainties for the default MET
   runMetCorAndUncFromMiniAOD(process,
                          isData=True (or False),
                         )

   # Now you are creating the bad muon corrected MET
    process.load('RecoMET.METFilters.badGlobalMuonTaggersMiniAOD_cff')
    process.badGlobalMuonTaggerMAOD.taggingMode = cms.bool(True)
    process.cloneGlobalMuonTaggerMAOD.taggingMode = cms.bool(True)

    from PhysicsTools.PatUtils.tools.muonRecoMitigation import muonRecoMitigation

    muonRecoMitigation(
                       process = process,
                       pfCandCollection = "packedPFCandidates", #input PF Candidate Collection
                       runOnMiniAOD = True, #To determine if you are running on AOD or MiniAOD
                       selection="", #You can use a custom selection for your bad muons. Leave empty if you would like to use the bad muon recipe definition.
                       muonCollection="", #The muon collection name where your custom selection will be applied to. Leave empty if you would like to use the bad muon recipe definition.
                       cleanCollName="cleanMuonsPFCandidates", #output pf candidate collection ame
                       cleaningScheme="computeAllApplyClone", #Options are: "all", "computeAllApplyBad","computeAllApplyClone". Decides which (or both) bad muon collections to be used for MET cleaning coming from the bad muon recipe.
                       postfix="" #Use if you would like to add a post fix to your muon / pf collections
                       )

    runMetCorAndUncFromMiniAOD(process,
                           isData=runOnData,
                           pfCandColl="cleanMuonsPFCandidates",
                           recoMetFromPFCs=True,
                           postfix="MuClean"
                           )


    process.mucorMET = cms.Sequence(                     
        process.badGlobalMuonTaggerMAOD *
        process.cloneGlobalMuonTaggerMAOD *
      #process.badMuons * # If you are using cleaning mode "all", uncomment this line
        process.cleanMuonsPFCandidates *
        process.fullPatMetSequenceMuClean
        )

   process.p = cms.Path(
                     process.mucorMET  *
                     process.fullPatMetSequence * # If you are re-correctign the default MET
                     process.yourAnalyzer
               )

vector                      "slimmedMETs"               ""                "PAT"  --> Out of the Box MET
vector                      "slimmedMETs"               ""                "PROCESSNAME" --> Out of the Box MET + Re-corrected for JEC
vector                      "slimmedMETsMuClean"        ""                "PROCESSNAME" --> Reclustered MET for the bad muons  + Re-corrected for JEC

How to re-correct MET based on e/gamma gain switch correction on the fly for Re-Miniaod Data (Release: CMSSW_8_0_X, X>=26_patch1)

For some portion of the events, the pf candidate referencing of the photons and electrons is not working. As a fall back solution, in the MET tool, a dR matching was implemented, however for a smaller subset of the events the dR matching was found to be too tight. If you want to recover this subset of events, one can follow the below recipe.

Full recipe is below Hide the recipe

   git cms-merge-topic cms-met:METRecipe_80X_part2 -u

   ## Following lines are for default MET for Type1 corrections.
   from PhysicsTools.PatUtils.tools.runMETCorrectionsAndUncertainties import runMetCorAndUncFromMiniAOD

   # If you only want to re-correct for JEC and get the proper uncertainties for the default MET
   runMetCorAndUncFromMiniAOD(process,
                          isData=True (or False),
                         )

   # Now you are creating the e/g corrected MET on top of the bad muon corrected MET (on re-miniaod)
   from PhysicsTools.PatUtils.tools.corMETFromMuonAndEG import corMETFromMuonAndEG
   corMETFromMuonAndEG(process,
                    pfCandCollection="", #not needed                                                                                                                                                                                                                                                                                                                      
                    electronCollection="slimmedElectronsBeforeGSFix",
                    photonCollection="slimmedPhotonsBeforeGSFix",
                    corElectronCollection="slimmedElectrons",
                    corPhotonCollection="slimmedPhotons",
                    allMETEGCorrected=True,
                    muCorrection=False,
                    eGCorrection=True,
                    runOnMiniAOD=True,
                    postfix="MuEGClean"
                    )
     process.slimmedMETsMuEGClean = process.slimmedMETs.clone()
     process.slimmedMETsMuEGClean.src = cms.InputTag("patPFMetT1MuEGClean")
     process.slimmedMETsMuEGClean.rawVariation =  cms.InputTag("patPFMetRawMuEGClean")
     process.slimmedMETsMuEGClean.t1Uncertainties = cms.InputTag("patPFMetT1%sMuEGClean")
     del process.slimmedMETsMuEGClean.caloMET
 
     # If you are running in the scheduled mode:
     process.egcorrMET = cms.Sequence(
        process.cleanedPhotonsMuEGClean+process.cleanedCorPhotonsMuEGClean+
        process.matchedPhotonsMuEGClean + process.matchedElectronsMuEGClean +
        process.corMETPhotonMuEGClean+process.corMETElectronMuEGClean+
        process.patPFMetT1MuEGClean+process.patPFMetRawMuEGClean+
        process.patPFMetT1SmearMuEGClean+process.patPFMetT1TxyMuEGClean+
        process.patPFMetTxyMuEGClean+process.patPFMetT1JetEnUpMuEGClean+
        process.patPFMetT1JetResUpMuEGClean+process.patPFMetT1SmearJetResUpMuEGClean+
        process.patPFMetT1ElectronEnUpMuEGClean+process.patPFMetT1PhotonEnUpMuEGClean+
        process.patPFMetT1MuonEnUpMuEGClean+process.patPFMetT1TauEnUpMuEGClean+
        process.patPFMetT1UnclusteredEnUpMuEGClean+process.patPFMetT1JetEnDownMuEGClean+
        process.patPFMetT1JetResDownMuEGClean+process.patPFMetT1SmearJetResDownMuEGClean+
        process.patPFMetT1ElectronEnDownMuEGClean+process.patPFMetT1PhotonEnDownMuEGClean+
        process.patPFMetT1MuonEnDownMuEGClean+process.patPFMetT1TauEnDownMuEGClean+
        process.patPFMetT1UnclusteredEnDownMuEGClean+process.slimmedMETsMuEGClean)

   process.p = cms.Path(
                     process.fullPatMetSequence * # If you are re-correctign the default MET
                     process.egcorrMET  *
                     process.yourAnalyzer
               )

vector                      "slimmedMETs"                 ""                "PAT"  --> Out of the Box MET
vector                      "slimmedMETsMuEGClean"        ""                "PAT"  --> Out of the Box e/gamma + muon corrected MET
vector                      "slimmedMETsMuEGClean"        ""                "PROCESSNAME" --> Reclustered MET for the bad muons + e/gamma + Re-corrected for JEC