Implementation of Marginal CPI with libSTREAM

    * go to liftandrelax/CPP
    * edit Makefile according to local machine
    * call make
    * go to liftandrelax/STREAM_Python
    * export PYTHONPATH=../CPP/:/PATH/TO/pymlns/
    * export LD_LIBRARY_PATH=../CPP/

mln2fg.py

    * ./src/mln2fg.py --cnf ./resources/smokers.mln
          o This will create a weighted CNF factor graph from the MLN called 'smokers_cnf.fg'
    * ./src/mln2fg.py --cnf ./resources/smokers.mln ./resources/smokers.db
          o This will create a weighted CNF factor graph from the given smokers.mln with respect to the evidence given in smokers.db
          o The evidence file accepts evidence in the form such as smokes(Anna), friends(Anna, Bob). Each evidence must be in a separate line.

runCbpOnFgFile.py

    * ./src/runCbpOnFgFile.py smokers_cnf.fg
    * This will run CBP on the smokers_cnf factor graph and output the marginal of the variables

runMarginalCpi.py

    * ./src/runMarginalCpi.py smokers_cnf.fg