WrCoffea

Coffea Introduction

To install coffea:

python3 -m pip install coffea --user

All coffea documentation is hosted at https://coffeateam.github.io/coffea/

Getting Started

Begin by cloning the repository,

git clone git@github.com:UMN-CMS/WrCoffea.git
cd WrCoffea

Running the analyzer

Example

The command below will locally analyze one root file from each of the 2018 UL background MC samples:

python3 run_analysis.py UL18_bkg allMC --hists example_hists.root --masses example_masses.root  --max_files 1

Two files are outputted: a root file containing histograms of kinematic variables across all basic analysis regions (example_hists.root), and a root file with branches of the 3-object invariant mass ($m_{ljj}$) and 4-object invariant mass ($m_{lljj}$) (example_masses.root).

Arguments

To run the analyzer, a sample set and process must be specified as arguments:

• Sample sets: Currently, only UL18_bkg has been created, but there are also plans to include UL18_signal and UL18_data sample sets, as well as the rest of Run II (2016 and 2017).
• Process: The MC process to be analyzed. Options for individual processes are DYJets, tt+tW, WJets, Diboson, Triboson, ttX, SingleTop, or to analyze the entire sample set, use allMC.

To analyze over all files for the given process, simply omit the --max_files flag. To run the analyzer without computing any output files (perhaps for debugging purposes), omit both --output_hists and --output_masses.

For more information, enter:

python3 run_analysis.py --help

Future Work

In the near future there are plans to scale out the framework to use distributed computing resources needed to analyze all 2018 UL MC samples and files.