DelphesTree_converter

Repository with python code to convert root files produced through Delphes into pandas DataFrame.

It's recommended to use python 3.10.4 to run the scripts, and the required python modules.

Conda instalation

If your python version is not thisone, you can get it with a conda environment. Check if you have conda installed by doing:

conda --version

If you see anything different than `conda: command not found``, you have conda. Otherwise, download it by doing

wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh

once downloaded, do

chmod +x Miniconda3-latest-Linux-x86_64.sh
./Miniconda3-latest-Linux-x86_64.sh

to install it. Follow the instructions untill the end. Once installed activate conda if you didn't activated it by default in the instalation process doing:

source /home/atehort/miniconda3/bin/activate

Now you may want to create a virtual environment for this task by doing:

conda create -n <name_environment> python=3.10.4

And activate it by doing

conda activate <name_environment>

Instalation of the requirements

When you create the environment you'll get a bunch of modules, the otherones needed you can get them by doing

pip install coffea==2025.3.0

Running the code

The main code is stored in dt_converter.py. It has a class called Converter that can be used as follows:

tree_test = Converter(fname)
tree_test.generate({"Jet": ["PT", "Eta", "Phi", "Mass", "BTag", "TauTag"],
                    "Muon": ["PT", "Eta", "Phi", "Charge"],
                    "Electron": ["PT", "Eta", "Phi", "Charge"],
                    "MissingET": ["MET", "Phi"]}, 
                    jet_elements = 4, e_mu_elements = 2)
df = tree_test.df

You may include in the branches from the Delphes object that you're interested for your analysis as the example states.

Use the split_events from ùtils.py to store the dataframe split it into .parquet/.coffea files discriminating by the number of jets per event as follows:

split_events(df, label = 'Signal_37', path = '/store/atehort/', file_type = 'coffea')

The runner.py script is thought to build different files from a rootfile, each one with a pd.DataFrame that includes the kinematic information of the first two muons and the jets depending on the jet multiplicity

Use the runner via terminal as:

python3 runner --path <Path_To_The_Root> --label <label_for_the_output_file> --output_path <path_to_the_output_dir> --file_type <desired_output_type>

Supported output types are parquet, coffea or csv

Any question, please write to [email protected]