ttalps

This GitLab repository contains the code so far accumulated in the exploration of signatures of ALPs produced together with a top antitop pair at the CMS detector at the LHC, with the ALP decaying into a muon pair. We use this code to compare the signatures of signal events (tta->ttmumu) with background events which do not have an ALP (ttmumu) to separate the two sets of events on the basis of their kinematic variables.

Setup of virtual invironment

• For running the code with all necessary libraries, run with python3 venv as: python3 -m venv ttalps. For invironment name ttalps, in the ttalps directory.
• Activate: source bin/activate
• Download libraries: pip install -r requirements.txt
• Deactivate deactive.

Structure of the code

The code is made up of four auxiliary python files:

• The file aux.py contains the input and output paths used in all Jupyter notebooks as well as the HiddenPrints context manager.
• The file data_classes.py contains the definitions of the classes for FourVector, FourMomentum, Particle, Event, and Dataset which we encapsulate our data in. They contain many useful methods for calculating observables.
• The file myplot.py contains convenient shortcuts to plots we use, as well as their colour scheme and a dictionary of commonly used labels for plot axes, dirLabels, for consistency.
• The file physics.py contains the used interface to TdAlps, functions for the calculation of lifetimes and branching ratios, as well as a dictionary of the physical constants needed, sm.

The package TdAlps used here comes from the Github repository https://github.com/TdAlps/TdAlps and is based on the paper hep-ph:[2012:12272] which describes Lagrangian we use and its RG-evolution. The package allows for the running of couplings in the ALP model we use.

We use these .py-files in the Jupyter notebooks:

• Decays.ipynb explores the ALPs lifetime and branching ratio in the considered coupling setup
• Data_Exploration.ipynb works with the samples to explore the generated events' parameter space
• dc_tests.ipynb is used to test the classes and methods of dataclasses.py with a reduced dataset

The folder MadAnalysis contains .cpp and .h files that are scripts for the program MadAnalysis5 which is used here to convert .hepmc files with generated events to .txt files which contain only the relevant variables for our analysis. In an Expert Mode analysis ALPanalysis.cpp in MadAnalysis5, these files would be placed in the folder ALPanalysis/Build/SampleAnalyzer/User/Analyzer/ and run through the command line. More information on the use of MadAnalysis5 can be found at https://madanalysis.irmp.ucl.ac.be/. The ALPanalysis.cpp file produces from a .hepmc file of ackground events several files:

• muon_count.txt which contains the number of muons and antimuons in every event
• muon_pair_count.txt which contains the number of muon-antimuon-pairs in every event
• muon_pair_parents.txt which contains the PDG-ID of every muon-antimuon-pair's parent in the dataset
• muon_data.txt which is the data to be read into the .ipynb files, containing
  • a list of all top quarks in the event (as we know there to be only one top quark in every event, this is the same top quark in different intermediate states; the last particle in the list is the 'final state' top)
  • a list of all anti-tops in the event (again, with the last in the list being the final state antitop)
  • a list of muon-antimuon-pairs (listed as muon, then antimuon)
  • a list of unpaired muons and antimuons
  • each of the particle's data is given by
    • its PDG-ID
    • its production vertex (given by the function decay_vertex)
    • its four-momentum (E, px, py, pz)
  • The particles' properties are separated by \t, the particles by |, and the events by \n.

Running MadAnalysis5

• Start by running MadAnalysis5 in expert mode ./bin/mg5 -e. Assuming the directory and workspace is named ALPanalysis:
• Change the ALPanalysis/Build/SampleAnalyzer/User/Analyzer/ accounding to ALPanalysis.cpp in the MadAnalysis folder
• Run source setup.sh
• In ALPanalysis compile with make
• Add the folder ALPanalysis/Output/TXT - otherwise the .txt files won't be saved while running the analysis
• Add the file input.txt including the path to the input background .hepmc file
• Run ./MadAnalysis5job input.txt

Running the .ipynb remotely with Jupyter notebook

• In lxplus start jupyter notebook with no browser jupyter notebook --no-browser --port=<port-number>. Specified port number is optional
• On local computer tunnel the remote local host with ssh -N -f -L localhost:8888:localhost:<port-number> <username>@lxplus<7xx>.cern.ch, for current lxplusnumber 7xx, and enter password
• In browser type localhost:8888 For VS code only first step is needed, then open browser under PORTS is VS code.