python - How to create libpgm discrete bayesian network CPD/data file from raw data -

i trying create bayesian network using libpgm library in python. documentation (http://pythonhosted.org/libpgm/#documentation) provides details of how use library given network structure , parameters learnt. unable figure out how create input dictionary network given here (http://pythonhosted.org/libpgm/unittestdict.html) . current have raw data in format of variables , observations. approach should take find network structure , parameters automatically? in advance

use pgmlearner automatically retrieve data looking https://pythonhosted.org/libpgm/pgmlearner.html