Please send us your bug reports!

At any point, you can use issue the command

marmosetHelp

if you are lost. It will give you a list of the relevant executables and options that you can use in the current directory. Most executables must be run from a valid runDir.

• Create a runDir, anywhere you want.

makeRunDir <runDirName>

• Put a target signal monte carlo file in the data/ directory. This can be an .lhco.gz file or a Pythia .pyt file.
• Convert signal monte carlo to MPotato signatures for later fitting.

processData <data/dataName>
processData --all

• Create an .oset file in the oset/ directory. (There should be some sample .oset models there already.) Read up on the OSET file format to learn more about rolling your own.
• Parse that OSET model and populate the relevant trees.

cd <path/to/runDirName>
parseOSET <osets/OSETName.oset>

• Generate MC for each of the trees.

generateProcess --all

If you only want to generate MC for the trees in a particular OSET (to save time if you have many OSETs defined), you can type

generateProcess --oset <OSETName>

• See below for additional options

• Run the fitter.

fitOSET <OSETName> <data/target.sigs>

If you have a multiprocessor machine, you can use the –multiProc n flag to use n processors at a time. Similarly, use –condor n or –PBS n to generate MC by sending n jobs to a Condor CAF or PBS batch system.

• -v flag will print much more output to screen than normal operation; use it to see what's gonig wrong.
• –once will only attempt to execute one step in the Monte Carlo generation, and can be useful for debugging.

generateProcess <process>

(where <process> is the name of a process (e.g. p000001), not a full path) will just run generateProcess on a single process