Starting from #749, but keeping also support for older electron IDs (Spring15 and later) and the custom working points we had before (e.g. from susy PAG).
The MVA ID discriminators are all evaluated using the VID backend, even when we apply a custom WP
The POG-provided MVA WPs are taken also from VID
Cut-based IDs for Fall17V2 are from VID, V1 there's both the VID implementation and our old ones.
I've validated the outputs against NanoAOD and the previous Heppy implementation, and found very good agreement (there's a few rare events where the MVA score changes in far decimal places, but I think this is not unexpected for a change of code)
Starting from #749, but keeping also support for older electron IDs (Spring15 and later) and the custom working points we had before (e.g. from susy PAG).
Based on CMSSW_9_4_12
The differences wrt to #749 can be better seen in this diff: https://github.com/gpetruc/cmg-cmssw/compare/merge_749...gpetruc:eleID_new_and_old
Important: add
/DataFormats/FWLite/to sparse checkout to avoid crashes.@cbernet @GaelTouquet @steggema @lucastorterotot