Update PU weight file for 2018

[bortigno]

Working on branch issue-57-update-up-2018.

Checking what I did for 2017 in the issue #6 to remind myself how to do it.

One small modification is that in 10_2_11_patch1 the pileupCalc.py script has a small bug (an extra parenthesis) in line 253, so I had to find where it was (which pileupCalc.py) then copy it to my working dir, fixing the bug and run it from my local dir, using the following options:

python pileupCalcLocal.py 
   -i data/JSON/Cert_314472-325175_13TeV_17SeptEarlyReReco2018ABC_PromptEraD_Collisions18_JSON.txt 
  --inputLumi=data/Pileup/PU_data_2018.txt 
  --calcMode=true 
  --minBiasXsec 69200 
  --maxPileupBin 99 
  --numPileupBins 99 data/Pileup/PU_data_2018.root

Where the data/Pileup/PU_data_2018.txt comes from https://cms-service-dqm.web.cern.ch/cms-service-dqm/CAF/certification/Collisions18/13TeV/PileUp/pileup_latest.txt The uncertainty use here is 2.5% following this twiki recommendation: https://twiki.cern.ch/twiki/bin/view/CMS/TWikiLUM#CurRec based on LUM-18-002. So I created the systematics files using the following cross sections:

69200 (1+0.025) = 70930 69200 (1-0.025) = 67470

python pileupCalcLocal.py 
  -i data/JSON/Cert_314472-325175_13TeV_17SeptEarlyReReco2018ABC_PromptEraD_Collisions18_JSON.txt 
  --inputLumi=data/Pileup/PU_data_2018.txt 
  --calcMode=true 
  --minBiasXsec 70930 
  --maxPileupBin 99 
  --numPileupBins 99 data/Pileup/PU_data_2018_up_v0.root

and

python pileupCalcLocal.py 
  -i data/JSON/Cert_314472-325175_13TeV_17SeptEarlyReReco2018ABC_PromptEraD_Collisions18_JSON.txt --inputLumi=data/Pileup/PU_data_2018.txt 
  --calcMode=true 
  --minBiasXsec 67470 
  --maxPileupBin 99 
  --numPileupBins 99 data/Pileup/PU_data_2018_down_v0.root

The MC scenario I took from GitHub: https://github.com/cms-sw/cmssw/blob/CMSSW_10_2_X/SimGeneral/MixingModule/python/mix_2018_25ns_JuneProjectionFull18_PoissonOOTPU_cfi.py

following the recommendation of this twiki: https://twiki.cern.ch/twiki/bin/view/CMS/PileupScenariosRun2

and used it to modify the script GeneratePU.py and ran it producing the file data/Pileup/PU_wgt_2018_Autumn18_v0.root

To do:

• [ ] Need to update Analyzer_cff with the new PU_wgt file
• [ ] Test it

[bortigno]

Found big stupid bug. I branched out from master instead of master_2018_102X. Fixed now merging master_2018_102X to the branch issue-57-update-up-2018. It seems that merge was easy without conflicts. Though I found another problem: the muon_SF json in the *_cff.py is still poiting to the 2017 version. Reopened #55.

Things seem to be not kosher in this branch, maybe the branch merge didn't work as well as I thought.

[bortigno]

Fortunately I discovered that I just forgot to "git pull" the master_2018_102X.

Now I tried to merge the master back to issue-57-update-pu-2018 but the automatic merge failed. I need to solve conflicts on python/Analyzer_data_cff.py and python/Analyzer_MC_cff.py

This is the distribution of the MC PU_wgt on a signal sample:

h1_pu_weight_2018.pdf

Don't see pathologies.

Comparing the MC nVertices distribution with the data PU disrtibution it is clear that it didn't work

Looking at the PU profile saved in Generate.py for the MC, it does not look similar to the one that I get looking at the nVerticies in the MC tuples:

I think the right thing to do is to compared nPU in the MC with the nVertices in the data. In that case there is a much better agreement.

[bortigno]

Merged in PR #64