Back Skip to the content.

Description

The gundamFitter app is an interface to the FitterEngine which is in charge of setting up datasets, samples, fit parameters according to a provided config file.

The list of options can be displayed by running gundamFitter without command line arguments. The program should immediately stop after complaining that no config file was provided.

Usage

Most inputs are defined in the fitter config file and are called using the -c argument followed by the path to corresponding configuration file.

gundamFitter -c path/to/config.yaml

Config files can be written in the .yaml format or .json. Alternatively one can use gundamConfigUnfolder to convert .yaml inputs into .json.

The number of parallel threads used by the fit can be defined by using the -t argument as follows:

gundamFitter -c path/to/config.yaml -t 15

If not specified in the config files one can also provide the path to the output file with the -o argument:

gundamFitter -c path/to/config.yaml -o path/to/output.root -t 15

Override files can be passed as arguments with -of to override the default config without having to modify it:

gundamFitter -c path/to/config.yaml -o path/to/output.root -of path/to/override.yaml -t 15

Trigger options

A certain number of trigger options can be used depending on what the user needs:

Option Description
-c, --config-file Specify path to the fitter config file
-of, --override-files Provide config files that will override keys
-O, --override Add a command line override [e.g. /fitterEngineConfig/engineType=mcmc]
Runtime Options Description
–pca Enable principle component analysis for eigen decomposed parameter sets
-t, –nb-threads Specify nb of parallel threads
-s, –seed Set random seed
Fit Options Description
-a, –asimov Use MC dataset to fill the data histograms
–skip-hesse Don’t perform postfit error evaluation
–toy Run a toy fit (optional arg to provide toy index)
–inject-parameters Inject parameters defined in the provided config file
Output Options Description
-o, –out-file Specify the output file
–out-dir Specify the output directory
–appendix Add appendix to the output file name
–scan Enable parameter scan before and after the fit (can provide nSteps)
–scan-line Provide par injector files: start and end point or only end point (start will be prefit)
–one-sigma Generate one sigma plots
–light-mode Disable plot
generation  
Debug Options Description
-d, –dry-run Perform the full sequence of initialization, but don’t do the actual fit.
-dd, –super-dry-run Only reads the config files.
-ddd, –hyper-dry-run Only unfolds the config files.
–debug Enable debug verbose (can provide verbose level arg)
–kick-mc Amount to push the starting parameters away from their prior values (default: 0)
–ignore-version Don’t check GUNDAM version with config request
-me, –max-events Set the maximum number of events to load per dataset
-fe, –fraction-of-entries Set the fraction of the total entries of each TTree that will be read

Additional Options

Option Description
–skip-simplex Don’t run SIMPLEX before the actual fit
–one-sigma Generate one sigma plots
–light-mode Disable plot generation
–no-dial-cache Disable cache handling for dial eval

For a complete list of options run command without arguments.

Config options

Option Type Description Default
fitterEngineConfig json FitterEngine config  
minGundamVersion string gundamFitter will stop if the version is lower  
outputFolder string Folder where the output file is written ./