A converter for JE derived datasets into the BerkeleyTree format.

Contact • Dependencies • Installation • Usage

Contact

Authors: Kirill Naumov, Tucker Hwang, Florian Jonas
Contact: Tucker via Slack or tucker_hwang@berkeley.edu

Dependencies

This package comes in three components: a downloader, a converter, and a scheduler tool.

• The downloader requires a valid Python 3 installation with the rich module installed, and the AliEn tools, in particular alien_find, alien_cp, and alien_ls.
• The converter depends only on ROOT and the yaml-cpp development package and is compiled with make. The name of the development package varies from system to system, but is typically called yaml-cpp-dev or yaml-cpp-devel.
• The scheduler, written in Bash, requires an accessible Slurm configuration. The jobs are scheduled via sbatch.

Installation

To build the converter binary, you can simply run make in the root of the repository or run make -C <path/to/root>. To compile in debug mode, run with make BUILD=debug. The binary is built into an executable called converter in the bin directory. To run the converter standalone:

./bin/converter [args]
  --input-filelist=<file>, -i <file>  : Text file list of input AO2D files to be converted
  --output-filename=<file>, -o <file> : Path to the output BerkeleyTree ROOT file (default: BerkeleyTree.root)
  --config-file=<file>, -c <file>     : YAML file containing cuts to be implemented on the converted data (default: tree-cuts.yaml)
  --create-histograms                 : Build QA histograms from the converted data
  --save-clusters                     : Save cluster information
  --is-mc                             : The data is produced from Monte Carlo Simulation (currently just a placeholder)
  -v, -vv                             : Increase verbosity; -v is INFO, and -vv or higher is DEBUG

Usage

For general usage and documentation, please refer to the README in the scripts directory. For specifics on how to use the converter on the NERSC system, please see the NERSC-specific instructions.

Future improvements

Check our issues page!