Instructions to Setup and Run Simulation

• Clone this repo: git clone --recursive [email protected]:RobMina/ldmx-sim-visdecay.git

• cd into the new directory: cd ldmx-sim-visdecay

• The container runner available on Rivanna is apptainer, but the ldmx-sw environment assumes we will be using either docker or singularity. So, we need to make an alias: module load apptainer alias singularity="apptainer"

• Set the TMPDIR environment variable (necessary for ldmx-env.sh to run the first time on Rivanna): export TMPDIR=/scratch/<your Rivanna username>

• Setup the ldmx-sw environment. This will take a long time (~30 minutes) the first time you do it, but will be relatively quick after the first time. source ldmx-sw/scripts/ldmx-env.sh

• Compile the updated code. This also takes a while, but only the first time. ldmx compile

• You are now ready to run the scripts listed below. The scripts/gen_unscaled_library.py script is the most direct way to create a MG dark brem vertex library, which is needed by G4DarkBreM to simulate dark brem interactions within the detector. The scripts/gen_signal_samples.py script is the most direct way to run the detector simulation using ldmx, and it requires that the MG library has already been generated.

• On Rivanna, a MG dark brem vertex library is in /standard/ldmxuva/data/dblib and some sample signal files are in /standard/ldmxuva/data/eat_vis_signal

Table of Contents

Scripts for Generating MadGraph Dark Brem Library

• scripts/db-gen-lib-env.sh: Set up the MadGraph dark brem container. This is sourced within scripts/run_db_gen_then_extract.sh. You can also source this to run the MadGraph generation interactively.

• scripts/run_db_gen_then_extract.sh: Set up, run MadGraph to generate dark brem vertices in an lhe file, then run g4db-extract-library to extract the relevant info (recoil e and A' kinematics) to a csv file. This should be run from the base directory of the repository. Because MadGraph sometimes fails to generate the desired number of vertices, this script automatically re-runs the generation with sequentially increasing run numbers until the desired number of vertices has been generated.

• scripts/gen_unscaled_library.py: Automatically generate batch scripts and call sbatch for specified materials, A' masses, and incident energies. This creates a library of unscaled DB vertices that is needed for using the G4DarkBreM module in simulation. Each material-mass-energy point is submitted as a separate job.

Scripts for Studying G4DarkBreM Dark Brem Scaling

• scripts/run_g4db_scale.sh: Set up and run G4DarkBreM scaling of dark brem vertices, starting with a csv file created by g4db-extract-library and producing another csv file using g4db-scale. Note that you must scale from a higher energy to a lower energy. This should be run from the base directory of the repository.

• scripts/perform_scalings.py: Automatically run G4DarkBreM scaling for a variety of materials, A' masses, and energy scaling points. This runs interactively rather than using sbatch.

• scripts/compile_dblib_into_df.py: Merge and compress csv files from gen_unscaled_library.py and perform_scalings.py into .feather files: one unscaled and one scaled for each material and A' mass.

• scripts/fill_dblib_scaling_hists.py: Create histograms from the .feather files containing unscaled and scaled dark brem vertices to validate the scaling done in G4DarkBreM.

• scripts/scaling_studies.ipynb: A Jupyter notebook with code to validate the scaling done in G4DarkBreM.

Scripts for Generating LDMX Dark Brem Signal Samples

• scripts/test_config.py: A config file passed to ldmx fire. Configures the generator to scale A' kinematics and decays using the updated G4DarkBreM module. Takes two arguments: the path to a MG dark brem library and the number of events to simulate.

• scripts/decay_validation.ipynb: A Jupyter notebook with code to validate the handling of A' decays within G4DarkBreM.

• scripts/setup_ldmx_and_fire.sh: setup, then call ldmx fire. Assumes that the config scripts takes the same arguments as test_config.py: namely, the path to a MG dark brem library and the number of events.

• scripts/gen_signal_samples.py: automatically configure and launch grid jobs using sbatch to generate signal samples for each mass point and run number. Corresponding MG dark brem libraries must have already been created and must be in csv format. You should modify the paths defined near the bottom of this script before using it.