kit-cms / artus Goto Github PK

For e/gamma there is no implementation for residual scale and smearings corrections available. This should be implemented. Naturally, I recommend to do this in the placeholder function AdditionalCorrections in the ElectronCorrectionsProducer.

Dear Daniel, all,

Is this branch still needed?

Or can we delete it?

https://github.com/KIT-CMS/Artus/tree/harry_workaround_segfault_ROOT6

Cheers,

Artur

Run 2 Rochester Corrections for UL campaign https://twiki.cern.ch/twiki/pub/CMS/RochcorMuon/roccor.Run2.v4.tgz need to be added into the code /Utility/interface/RoccoR.h and /Utility/src/RoccoR.cc.

Furthermore, it is not necessary anymore to differentiate the treatment of 2016 Rochester Corrections.

Also, there might be some changes in the recommended methods to compute the scaleFactors, since the methods

//old, should only be used with 2017v0
double kScaleFromGenMC(int Q, double pt, double eta, double phi, int n, double gt, double w, int s=0, int m=0) const; 
double kScaleAndSmearMC(int Q, double pt, double eta, double phi, int n, double u, double w, int s=0, int m=0) const;  
double kScaleFromGenMCerror(int Q, double pt, double eta, double phi, int n, double gt, double w) const; 
double kScaleAndSmearMCerror(int Q, double pt, double eta, double phi, int n, double u, double w) const;  

provided are deprecated and still in use in /KappaAnalysis/src/Producers/MuonCorrectionsProducer.cc.

For a handfull of data skims, it seems that for a portion of events at the beginning of the file the onLumi code in ElectronCorrectionsProducer is skipped and executed only for the next portion of events. In this call the objects in the Lumi tree of the file are bigger than they should be (presumably integer multiple), which leads to assertion error in the getIdNamesMap function.

It would be much nicer, if the variations would be separately saved as weights in the NTuple. This would only need adjustments in the postprocessing rather than producing new dedicated NTuples.

Currently there are no Rochester Corrections provided by CMS in the official TWiki. As soon as they are available the Rochester Corrections should also be applied for UL18 and UL16 datasets in a similar fashion to #15.