Sometimes, non-physical samples are produced from your sampler and stored in your result file. In order to fix this, the summaryclean executable removes these non-physical points from your result file, derives other possible posterior samples (using the conversion module), and saves the cleaned data as either a dat, hdf5 or json file ready to use with PESummary.


Most PESummary executables run summaryclean by default

To see help for this executable please run:

$ summaryclean --help
usage: summaryclean [-h] [-w DIR] [-s SAMPLES [SAMPLES ...]]
                    [--labels LABELS [LABELS ...]]
                    [--file_format {dat,lalinference,bilby,lalinference_dat}]

This executable returns a cleaned data file

optional arguments:
  -h, --help            show this help message and exit
  -w DIR, --webdir DIR  make page and plots in DIR
  -s SAMPLES [SAMPLES ...], --samples SAMPLES [SAMPLES ...]
                        Path to posterior samples file(s). See documentation
                        for allowed formats. If path is on a remote server,
                        add username and servername in the form
                        {username}@{servername}:{path}. If path is on a public
                        webpage, ensure the path starts with https://. You may
                        also pass a string such as posterior_samples*.dat and
                        all matching files will be used
  --labels LABELS [LABELS ...]
                        labels used to distinguish runs
  --file_format {dat,lalinference,bilby,lalinference_dat}
                        Save the cleaned data in this format

Therefore a biproduct of this executable is to enable the user to switch between file formats easily from the command line.