Monte Carlo simulation must be corrected using weights to provide predictions that can be compared to data. These corrections include generator-level event weights to account for NLO (and higher) calculations as well as scaling the simulation to the recorded data integrated luminosity.
The Generator-Level Analysis block handles these MC event weights as well as handling
bookkeeping of event selection cuts for each systematic variation. It can be
called using GeneratorLevelAnalysis.
To set up cut bookkeeping and generator-level analysis in your job, include the following text in your config.yaml file:
# Add the generator-level analysis block here
GeneratorLevelAnalysis: {}
Rerun your code to make sure that everything is working properly.
You’ll notice that there is a lot more output now from the CutBookkeeper,
which shows you that it is tracking many cutflows. Each one of these corresponds
to a weight in the generated event — usually to a variation of the PDF
(parton distribution function), scale, or some other parameter in event generation.
When you are satisfied with the code changes, commit and push them.