peircebayes (referred to as PB in this documentation) is a probabilistic logic programming language designed for inference in probabilistic models with Dirichlet priors. See¹ for more details.

PB is implemented in Prolog and Python (2.7).

Python core dependencies

numpy>=1.9.2
scipy>=0.15.1
matplotlib>=1.4.3
cython>=0.22.1

Other dependencies

For installation, see the hyperlinks of the dependecies.

The Prolog implementation used is yap. A standard YAP installation is fine.

The code for sampling uses the GNU scientific library ( gsl ).

pyCUDD is used for compiling BDDs in Python. If you're feeling lucky, try the steps below (otherwise see source files):

1. in the cudd folder (make sure the GCC flags are appropriately set in Makefile):

make clean
make
make libso

2. in the pycudd folder:

make depend
make

3. in your configuration file (e.g. ~/.bashrc) (replace the path to pycudd2.0.2 with your own):

export PYTHONPATH=$PYTHONPATH:~/Programs/pycudd2.0.2/pycudd
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:~/Programs/pycudd2.0.2/cudd-2.4.2/lib

PB install

To install PB (from PB folder):

sudo python setup.py install

Then try:

peircebayes -h

And tests (from PB folder):

cd tests
python tests.py core
python tests.py opt

For the moment, please cite PB using bibtex:

@misc{Turliuc15,
    title = {peircebayes - Probabilistic Abductive Logic Programming using Dirichlet Priors},
    author = {Calin Rares Turliuc},
    url = {http://github.com/raresct/peircebayes}
}

A paper describing PB has been accepted at the Second Workshop on Probabilistic Logic Programming. We will update the citation once we have the bibtex reference to the paper.

If you have any questions/problems/suggestions, don't hesitate to contact me at: raresturliuc [at] gmail [dot] com. Note that my willingness to help with installation issues is significantly reduced on OSs different from Ubuntu 14.10.