Boosted b-tag validation analysis
To set up:
Follow the directions to clone cms-btv-pog/RecoBTag-PerformanceMeasurements at https://github.com/cms-btv-pog/RecoBTag-PerformanceMeasurements
After setting up the PerformanceMeasurements package:
cd CMSSW_8_X_X/src
(Replace X's with the release number you are using)
git clone -b boostedbb_SFComm_80x_noSlimmedTrees git://github.com/cms-btv-pog/RecoBTag-BTagValidation.git RecoBTag/BTagValidation
scram b -j8
cd RecoBTag/BTagValidation/test/
cmsRun btagvalidation_cfg.py
To run systematics: src/BTagValidation.cc has b fragmentation, c fragmentation, ntracks, c->D fragmentation, K0/Lambda, and pileup up and down systematics implemented. For b fragmentation, in test/btagvalidation_cfg.py, doBFrag=True and you need to run once with doBFragUp=True and once with doBFragDown=True. For c fragmentation, doCFrag=True. For c->D fragmentation, doCDFrag=True. For K0/Lambda, doK0L=True.
For ntracks reweighting, you have to run test/dataMC_ntracks_reweight/nTracks_wt.py or test/dataMC_weight_calc/datamc_weight_calc.py over a file that already contains the QCD and data distributions, in order to get a root file that will reweight the QCD ntracks distribution to look like the data ntracks distribution. That file is fed to test/btagvalidation_cfg.py as FileNtracksWt and you must set doNtracksReweighting=True.
For pileup up and down, you use appropriate up/down PU Profile by setting File_PUDistData in test/btagvalidation_cfg.py properly. (usually +- 5% nominal minBias xsec)