============== Analysis_CMS ============

Tools for analyses of data from the CMS detector. The tools here are designed for a CMS based analysis, but do not depend on the CMSSW framework. Rather it relies mostly on having CERN ROOT. Below is a description of the directory contents and how they can be used.

-------------- ./fcncAnalysis --------------

Contains analysis code specifically for flavor changing higgs analysis. The analysis code is a ROOT TSelector which is designed for looping over a set of ROOT ntuples (TTrees in our case). The main piece of code is fcncAnalyzer_Template.C which is run with the following command,

./fcncLocal.csh

where is to identify the sample you are running over (creates a directory in the output root file), specifies the trigger stream that is desired (choice are electron, muon, mueg, or mc; mc doesn't apply any trigger selection), and period is either 2011 or 2012 (currently only works for 2012). This will output a number of histograms to the directory 'histos'. The script fcncLocal.csh is used to replace some information in the analysis code and execute the run.C file.

Additionally, there are some python scripts for submitting jobs to the lpc batch system. These allow you to select which datasets to run over and how you would like to split up the jobs for each data sample. The executable is batch_cfg.py. The class for batch submission is BatchMaster.py, and the file execBatch.csh is needed to configure the batch staging environment.

After the batch_cfg.py file has been modified to run over the desired datasets, it can be run with the following command

./batch_cfg.py

Both of the arguements are based on how the datasets have been defined in the config. By default, is assumed to be either 2011 or 2012 and the list of data is 'data', 'bg', or 'signal'.

-------------- ./fakeAnalysis --------------

Similar in structure to fcncAnalysis, the fakeAnalysis code is based off of the ROOT TSelector class. It is simpler and only requires that one execute the run.C file in ROOT as follows,

root -l run.C

The output is a set of histograms for determining the probability of leptons being faked by objects of the users definition. Definitions of these objects needs to be done in the relevant sections of the Selector class described briefly below.

----------- ./interface -----------

Contains header files for a set of container classes for physics objects. The base class is TCPhysObject which inherits from the ROOT TLorentzVector(http://root.cern.ch/root/html/TLorentzVector.html). This allows for saving kinematic variables and values that are common to various of physics objects (leptons, jets, etc.). The current set of classes are

• TCJet
• TCMuon
• TCElectron
• TCTau (needs additional development)
• TCMet
• TCTriggerObject
• TCPhoton
• TCGenParticle
• TCGenJet

All classes inherit from TCPhysObject with the exception of TCMet which inherits from TVector2 (http://root.cern.ch/root/html/TVector2.html).

----- ./src -----

Source files for container classes.

--------- ./plugins ---------

Contains definitions of classes that are helpful in general for physics classes. This includes both header and source files of the following classes,

• HistManager: Allows for filling and booking of histograms in a single line. Thanks, Andy.
• Selector: For selecting physics objects (electrons, muons, jets, etc.). Carries out identification of leptons based on a set of working points and produces a set of selected objects which can be retrieved from the instantiation of the Selector class.
• TriggerSelector: Used for identifying whether a given event has passed the trigger requirement. The trigger status is stored in 64-bit unsigned integer so there's some extra unpacking to figure out if a given trigger (specified by a string) has passed :|
• WeighUtils: Used for applying weights to a supplied event. This is done based upon various event variables, e.g.the kinematics of the leptons.
• EGammaMvaEleEstimator: a tool for applying BDT based identification of electrons. This is passed to the selector class.

--------- ./scripts ---------

A variety of scripts for doing such tasks as making overlays of histograms, making tables, simple fits, producing efficiency plots, etc.

For producing plots comparing background, data, and signal one can make use of the PlotProducer.py class. The working configuration for this is plot_cfg.py. This is really most useful if you have many plots with many background components because it requires a lot of configuration.

For producing tables the TableMaker class is useful. This is written with producing tables displaying the cut flow for a given analysis, but can be configured to output the entries of a histogram into nice LaTex or Twiki formatted tables.