Configuring GUNDAM

GUNDAM is configured using YAML (Json) input files that describe and control every aspect of the job. For the most part, the input files are directly handled as standard YAML, with one minor exception to include files into the configuration.

Best practices: Including other files into the configuration.

The recommended practice is to break the configuration into several input files that reflect the logical structure and to include them using the include file mechanism.

The GUNDAM libraries start by reading the entire configuration in to memory, and then it is interpreted into a json structure according to the usual rules with one exception. If the value of a string is found to end in the extension “.yaml”, then the string is treated as a filename relative to the current working directory to be included at that point, and the contents of the file is substituted into the input stream. For instance

fitterEngineConfig: "./inputs/engineConfiguration.yaml"

will read the contents of the “./inputs/engineConfiguration.yaml” file as a Json object add it to the full Json object as a child of the “fitterEngineConfig” field. This provides convenient ways to factor the full configuration into components.

minGundamVersion: 2.0.0
fitterEngineConfig:
    minimizerConfig:
        type: RootMinimizer
        minimizer: Minuit2
    likelihoodInterfaceConfig:
        jointPropabilityConfig:
            type: BarlowLLH

could be broken into three files:

topConfig.yaml

minGundamVersion: 2.0.0
fitterEngineConfig:
  minimizerConfig: "./minimizerConfig.yaml"
  likelihoodInterfaceConfig: "./likelihoodConfig.yaml"

minimizerConfig.yaml
```
type: RootMinimizer
minimizer: Minuit2
```

likelihoodConfig.yaml

jointPropabilityConfig:
type: BarlowLLH

Top Level Configuration Options.

Option	Type	Description	Default
fitterEngineConfig	json	FitterEngine config
minGundamVersion (the format is “v.r.p”)	string	gundamFitter will stop if the version is lower
outputFolder	string	Folder where the output file is written	./