Clean the GET Data with the Hough Transform Code

Implementation

This step in the analysis should be performed on an HPC. The script for the clean_events Hough transform cleaning code is in the main pytpc repository (pytpc/bin/clean_events) and the implementation in the Python module pytpc.cleaning. The clean_events script parses through a data file and removes noise.

This code:

  1. Does a nearest neighbor comparison to eliminate statistical noise
  2. Does a circular Hough transform to find the center of the spiral or curved track in the micromegas pad plane
  3. Does a linear Hough transform on (z,r*phi) to find which points lie along the spiral/curve
  4. Writes points and their distance from the line to an HDF5 file

The “Cleaning parameters” in the config file decide how aggressively the data is cleaned.

Usage

clean_events can be used as follows:

clean_events [-h] [--canon-evtids CANON_EVTIDS] config input output
                  # arguments in [] are optional

The optional CANON_EVTIDS arguments provides a path to an HDF5 file containing information to correct for misaligned CoBo event IDs.

The config argument takes the path to the proper config file. The input argument takes the path to the HDF5 file to be cleaned. The output argument takes the path to where the cleaned HDF5 file will be written and requires a filename such as clean_run_XXXX.h5 to create a file by that name. If no path, only a filename, is provided, the file will be created in the same directory as the original HDF5.

The following command will display this information in the terminal:

clean_events -h