search

cms-opendata-analyses / PhysObjectExtractorTool

This repository has working code examples (snippets) on how to access different physics objects in the context of CMSSW software.
6 stars 36 forks source link

2011 branch: updates needed (other than the poet config) #85

Open katilp opened 1 year ago

katilp commented 1 year ago
jmhogan commented 1 year ago

@katilp unfortunately I think we are kind of stuck on the JetResolution input file... we got this file from the Workbook instructions, and it's definitely for 2012: https://twiki.cern.ch/twiki/bin/view/CMSPublic/SWGuideJetResolution I'm not sure why this was never moved from this developer's cmssw fork to the main one.

Looking at the same folder in central cmssw, the 53X branch only contains resolutions marked as "Spring10": https://github.com/cms-sw/cmssw/tree/CMSSW_5_3_X/CondFormats/JetMETObjects/data Seems like there were no further updates to the contents of this directory.

Similarly, that TWiki seems to have been updated directly from 2010 to 2012. Revision 5 looks to be for 2010 data: https://twiki.cern.ch/twiki/bin/view/CMSPublic/SWGuideJetResolution?rev=5 And the next revision was not made until 2014.

So the information for 2011 looks to be a bit lost... Options:

katilp commented 1 year ago

OK, thanks for investigating!

I would expect JER for 2011 to be more recent than those for 2012 (because of their later reprocessing) but that does not change what you've found (or did not find...). They should still be 5_3_X.

I can try to do some further data archeology with that file JER file name, and then we can decide.

katilp commented 1 year ago

For the record going through JEC/JER meetings at that time:

And for the record, in general on that reprocessing (but no information on JER):

https://twiki.cern.ch/twiki/bin/view/CMSPublic/SWGuideFrontierConditions?rev=609#7_TeV_MC_Legacy_reprocessing_in

katilp commented 1 year ago

There's a collection of different tables for JER in https://twiki.cern.ch/twiki/bin/viewauth/CMS/JetResolution

and, in particular, for 2011

e.g

image

image

However, I'm not sure how these numbers correlate with those in the file JetResolutionInputAK5PF.txt that have function parameters and narrower eta bins.

jmhogan commented 1 year ago

Yes, the JER factors themselves are all accessible for each year, and I believe/hope we already have 2011 and 2012 different from each other.

But to apply them one needs to know the measured pT resolution of jets. That's the information stored in that text file, and that's what we're missing for 2011.

Julie

On Wed, Nov 23, 2022, 5:30 AM Kati Lassila-Perini @.***> wrote:

There's a collection of different tables for JER in https://twiki.cern.ch/twiki/bin/viewauth/CMS/JetResolution

and, in particular, for Run-1 https://twiki.cern.ch/twiki/bin/viewauth/CMS/JetResolution#History_of_JER_measurements

e.g. for 2011:

[image: image] https://user-images.githubusercontent.com/5859918/203533983-cabf4745-bb9d-4240-bc0e-f43ef5d670d3.png

However, I'm not sure how these numbers correlate with those in the file JetResolutionInputAK5PF.txt https://github.com/cms-opendata-analyses/PhysObjectExtractorTool/blob/2012/PhysObjectExtractor/JEC/JetResolutionInputAK5PF.txt that have function parameters and narrower eta bins.

— Reply to this email directly, view it on GitHub https://github.com/cms-opendata-analyses/PhysObjectExtractorTool/issues/85#issuecomment-1324912127, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACVFOJBBHAA3ULNUJB2KBKDWJX55PANCNFSM6AAAAAASBFMGVA . You are receiving this because you commented.Message ID: @.*** .com>

jmhogan commented 1 year ago

B-tagging efficiency setup is updated for 2011, separate from 2012.

https://github.com/cms-opendata-analyses/PhysObjectExtractorTool/commit/6ba3f381456c0ea8168ed23be09840837b953a44

katilp commented 1 year ago

Chat with @kirschen :

  1. JER scale factors

    Update those in the end of the jet analyzer i.e.

    std::vector<float>
JetAnalyzer::factorLookup(float eta) { //used in jet loop for JER factor value
  //eta input is > 0
  if(eta > 3.2) return {1.056, 0.865, 1.247}; // {factor, factor_down, factor_up}
  else if(eta > 2.8) return {1.395, 1.332, 1.468};
  else if(eta > 2.3) return {1.254, 1.192, 1.316};
  else if(eta > 1.7) return {1.208, 1.162, 1.254};
  else if(eta > 1.1) return {1.121, 1.092, 1.15};
  else if(eta > .5) return {1.099, 1.071, 1.127};
  else return {1.079, 1.053, 1.105};
}

    with the values in the table in https://twiki.cern.ch/twiki/bin/viewauth/CMS/JetResolution#JER_Scaling_factors_and_Unce_AN2 e.g.
    image

  2. For JetResolutionInputAK5PF.txt:

    This was the first that file was produced. It was on 8 TeV CMSSW_5_3_X MC. 2011 data was re-reco'ed with 5_3_X, and these values were not re-evaluated specifically for 8 TeV. The best approach is to use these same values for 2011

    Page 11 in these slides shows the values in the file for a certain pt (not mentioned)

katilp commented 1 year ago

For the future Run2 jets

Corrections are applied but smearing is still expected to be done at the mini/nano level because it involves random numbers so cannot be "undone" (although discussions were ongoing whether to do it for Mini/Nano and store the smearing factor used )

jmhogan commented 1 year ago

@katilp Aha, I was wrong to think we had already used the 2011 numbers in the 2011 branch... Fixed now!