This code requires python and the coffea package (link).
To setup python and coffea, follow the steps here (tested on CERN lxplus):
Setup python (steps copied from here). Download and install the most recent version of Miniforge:
wget https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-Linux-x86_64.sh
bash Miniforge3-latest-Linux-x86_64.sh
Activate miniforge
source /home/user/XXX/miniforge3/bin/activate
Install coffea into a separate environment.
conda install -n coffea coffea
Code may require additional python packages, install them via
conda install -n coffea [PACKAGE_NAME]
After each login, activate the miniforge installation and the coffea environment:
source /home/user/XXX/miniforge3/bin/activate
conda activate coffea
The code aims to dump the inputs for an anomalous jet search and save them to parquet files. These files can be opened with the python package awkward. An example file can be found here. Information from the file can be loaded as demonstrated in the following:
import awkward as ak
test = ak.from_parquet("test_QCD.parquet")
print(test.fields) # to get all properties stored in the array
feature = test.rho
The awkward arrays can also be converted to NumPy arrays:
print(feature)
array = ak.to_numpy(feature)
The QUAK training repository can be found here.