Crystalball currently supports prediction from a WSClean list of delta and Gaussian components with (log-) polynomial spectral shape (https://sourceforge.net/p/wsclean/wiki/ComponentList) into the MODEL_DATA column of a measurement set. This is done largely based on code available in the https://github.com/ska-sa/codex-africanus library.
Crystalball depends on python-casacore which builds from source. The dependencies mentioned at the following links must be installed in order for the build to succeed:
- https://github.com/casacore/casacore#building-from-source
- https://github.com/casacore/python-casacore#from-source
Create and activate a Python3 virtual environment
virtualenv -p python3 <name-of-virtualenv>
(systems without a Python2 installation don't even need the -p python3 specifier, but it doesn't hurt.)
source <name-of-virtualenv>/bin/activate
Pip install crystalball
pip install <path to crystalball>
Activate the virtual environment where you installed codex-africanus, xarray and xarray-ms (see above)
source <name-of-virtualenv>/bin/activate
Run crystalball
crystalball <file.ms> [-h] [-sm SKY_MODEL] [-rc ROW_CHUNKS] [-mc MODEL_CHUNKS] [-f FIELD] [-mf MEMORY_FRACTION] [-w REGION_FILE] [-po] [-ns NUM_BRIGHTEST_SOURCES] [-j NUM_WORKERS] [-o OUTPUT_COLUMN]