Files for Creating Jet Energy Correction Plots

The raw jet kinematics and observables of our files can be viewed in the raw_jets.ipynb notebook. This file is then scaled-up using Coffea in raw_jets_processor.ipynb, where a pkl file containing all the histograms of our dataset is produced.

Next we apply cuts/masks to our analysis, which can be seen in the qcd_jets.ipynb notebook; the goal of this file is to obtain a $p_T$ response curve. As before, this is scaled-up to include all our dataset qcd_jets_processor.ipynb.

The qcd_jets_processor.ipynb notebook produces histograms using the ROOT files in /samples, and corrects them using the pileup weights found in /data; the histograms are then dumped into pkl files.

The notebook jer_computations.ipynb uses the pkl files produced in qcd_jets_processor.ipynb to produce $p_T$ response curves, and then fits them to gaussian functions. The means and widths, and their respective errors are then put in a CSV file for analyzing.

Lastly, the jer_plotting.ipynb file is used to plot the JER curves as functions of the jet $p_T$ for all datasets.

Running the Files

Running on CoffeaCasa

To run on CoffeaCasa, simply access the site coffea-casa and sign in using either your x-certificate or CERN single sign-on credentials.

Running on LPC

For running on LPC, we follow the instructions from lpcjobqueue. First SSH to cmslpc by replacing <username> with your credential:

kinit <username>@FNAL.GOV
ssh <username>@cmslpc-sl7.fnal.gov -L 8888:localhost:8888

Next we must obatin a voms ticket, make a working area, and clone this repository to that area:

voms-proxy-init -voms cms
mkdir working_directory
cd working_directory
git clone https://github.com/cebarros/Jet_Analysis.git

We can now run the following commands to create a Singularity apptainer shell with a Coffea 0.7 environment:

curl -OL https://raw.githubusercontent.com/CoffeaTeam/lpcjobqueue/main/bootstrap.sh
bash bootstrap.sh
rm bootstrap.sh
./shell coffeateam/coffea-dask-cc7:0.7.22-py3.10-g7f049
jupyter lab --no-browser --port 8888

If successful, you should have an output with a URL simlar to:

http://localhost:8888/?token=...

Copy this URL and paste it into your browser to start the analysis. Once this has all been set up, you only need to run:

./shell coffeateam/coffea-dask-cc7:0.7.22-py3.10-g7f049
jupyter lab --no-browser --port 8888

To resume if disconnected or logged out.