This repository contains code to translate the HDF5 files output by ndlar_flow to Supera format for use by the DUNE machine learning reconstruction chain, lartpc_mlreco3d.
flow2supera depends on edep2supera, SuperaAtomic, larcv and h5flow. Install each of those repositories using the instructions on their respective READMEs and ensure that you can import them in python. Make sure the installation follows this order: larcv -> SuperaAtomic -> edep2supera -> flow2supera.
Once the prerequisites are met, simply run this command from the top directory:
python3 -m pip install .
The main executable script is located at bin/run_flow2supera.py relative to the top directory. The required arguments are the input and output file names and the configuration:
python3 bin/run_flow2supera.py -o <output_file> -c 2x2 <input_ndlar_flow_file>
Configuration keyword or a file path (full or relative including the file name). Supported configurations: 2x2, 2x2_data, mod1_data, 2x2_mpvmpr.
You can also specify the following optional arguments:
-nor--num_events: Number of events to process.-sor--skip: Number of first events to skip.-lor--log: Name of a log file to be created.
Upon successful completion, this will produce an output larcv-format file that can be used as input to the machine learning reconstruction.
Please read the contributing.md file for information on how you can contribute.
Distributed under the MIT License. See LICENSE for more information.