Customisation functions for jet tagging

[thesps]

Following from the recent discussions on centralising some jet tagging ntupling, here I add the methods to save the content we used for some previous studies.

Add variables from jet constituents

• pt
• eta
• phi
• mass
• pdgId

Save GenJet hadron&parton-based flavour (https://twiki.cern.ch/twiki/bin/view/CMSPublic/SWGuideBTagMCTools#Hadron_parton_based_jet_flavour)

[jmduarte]

@thesps this is great!

I was trying to do something similar, but my approach was going to be to save all the particle candidates using saveCands (like we do for ML MET training), but adding an index to tell it which jet it belongs to.

This has the benefit that you can save all the candidates, but then you have to do some postprocessing to make the training arrays.

@violatingcp @ddiaz006

[artlbv]

hi folks!
@thesps does the output ntuple you are pushing here correspond to what we have been using for the jet ID studies last year?
I was also wondering what ntuples the b-tag team @violatingcp was using if not this repo?

@jmduarte would it be easy to create training datasets with the format single jet+constituents from the files you were using?
The advantage of the method Sioni uses is that this is much closer to what the FW will be doing, right?

[artlbv]

maybe I can add @EmyrClement here to the thread