Useful scripts to analyze the ntuples produced with nano-AOD-tools (https://github.com/ttXcubed/nanoAOD-tools) The ntuples can be found in /nfs/dust/cms/user/gmilella/ttX_ntuplizer, separated per bkg, sgn, data samples and per years. The analysis scripts uses ROOT DataFrame (https://root.cern/doc/master/classROOT_1_1RDataFrame.html). In this repository, some examples to study AK8 variables are given. The scripts works per single file and includes:

• parsing of input file (importing x-section and MC genweight depending on the processes)
• creation of the output file
• creation of additional columns (not present in the file TTree) and event filtering
• histogram making which are then saved in the output file

A simple test on a single file can be done using the following command:

python ak8_multiplicity.py --input_file FILE --output_dir OUTPUT_DIR --year YEAR

where FILE can be fetched from the aforementioned repository (e.g. /nfs/dust/cms/user/gmilella/ttX_ntuplizer/bkg_2018_hotvr/merged/tt_dilepton_MC2018_ntuplizer_5_merged.root). OUTPUT_DIR and YEAR are choosen by the user. Additional option that can be used are:

• is_sgn or is_data
• sys: specifying the systematic (and the up/down variations) for which doing the analysis. The various systematics can be read from the ntuples

N.B.:

• the script fetches the name of the process from the file (e.g tt_dilepton) which is then used to fetch its cross-section in the xsec.yaml. Adjust the file name accordingly!
• in case of bkg files, I have split the same processes in multiple files to make the analysis faster. Therefore in the repository where the ntuples are fetched, there is an additional file sum_gen_weights.yaml that sums up the genweights for all the split files per process.
• the scripts (current version: commit 54c73fd) select event with exactly 2 opposite sign leptons, at least 2 AK4 (2b) to be outside the HOTVR jet cone (if the latter is present in an event)

Parallel analysis of files can be done by sendind as many condor jobs as ntuples. This can be done with the following commands:

python ttX_analysis_submission_template.py --OUTPUT_DIR --year YEAR
condor_submit ttX_analysis_condor_submission_new

additional arguments are is_sgn or is_data. The first command generates the ttX_analysis_condor_submission_new and the executable file depending on the arguments given. The ntuples are fetched automatically in the ttX_analysis_submission_template.py. You would need to modify the path according to where your ntuples are stored