Skip to content

Introducing ntuple skimming to bbll analysis

Giuseppe Callea requested to merge gcallea/easyjet:gcallea-bbll-ntupskimming into main

We noticed that our easyjet ntuples were huge compared to the ones from other analyses and thanks to the HHfw coordinator we realised that we were not using any skimming.

This MR:

  • Introduces the skimming in the HHbbllSelectorAlg scripts (Trigger, exactly 2 leptons and exactly 2 b-jets)
  • Gets rid of SelectionFlagsAlg (every requirement is now moved into HHbbllSelectorAlg)
  • Adds the bypass option to produce unskimmed ntuples

We validated the new implementation, achieving ~96% size reduction in a signal sample, and compared some distributions from the unskimmed vs skimmed samples. A summary of our checks can be found in the file in the attachmentHH-_bbll_ntuple_skimming.pdf.

The proposed skinning has been approved by the analysis team. The analysis README has been updated to account for the new changes

Edited by Giuseppe Callea

Merge request reports