GENIE Model Studies

A. Mastbaum [email protected], 2018/02

This repository contains scripts for comparing different models implemented in the GENIE neutrino event generator.

Production Scripts

There are production scripts in jobs, set up to generate event samples on the Fermilab grid. If using these, be sure to adjust output directories to suit your needs! Scripts are:

gevgen_models_flat.sh

    Ar40, choose nue/numu
    Flat flux, 0-3.5 GeV

gevgen_models_mono.sh

    Ar40, choose nue/numu
    Monoenergetic beam at 10 GeV

gevgen_models_ubflux.sh

    Ar40, choose nue/numu
    BNB-like flux (provided in data; place it somewhere in pnfs)

submit_gevgen_models

    Submit jobs for one of the event generation scripts, for all the GENIE
    model combinations.

The gevgen scripts take three arguments:

./gevgen_models_(type) [NVENTS] [CONFIG] [GENERATORS]

NEVENTS - Number of events to generate
CONFIG - A GENIE configuration (e.g. DefaultPlusMECWithNC)
GENERATORS - Specifies the generator list (e.g. Default+CCMEC+NCMEC)

Note that in order to enable a particular configuration, the cross section tables and physics list are read from a specified file location. In the current versions of the scripts, it reads and unpacks a tarball with "minified" data (Ar40 num/nue only) and settings for each configuration, from persistent pnfs (dCache).

Analysis Scripts

An analysis script named ggst reads in GENIE GST (summary tree) ntuples and plots of kinematic parameters, depending on the enabled generators (e.g. CCQE, CCMEC, all events). To build the ggst program, run make.

The basic workflow is to generate event samples incusively (e.g. Default+CCMEC+NCMEC) and with single generators enabled (e.g. CCMEC only) and feed those into ggst, which will compute for example an overall q0/q3 distribution for all events, and the joint proton kinetic energy distribution for CCMEC events with two-proton final states. There are a couple of additional scripts which take the per-generator output and condense it into summary plots.

There are two input modes, for reading a set of files as they come out of the production scripts (i.e. 1000 .gst.root files) or one file where they are hadd-ed together. In either case, the configuration and generator names are divined from the file path/name, which makes this somewhat fragile. See the file loading section of ggst.cpp for complete details.