Pybliss-Dockerfile

Synopsis

Bliss is a (great) tool to check wether or not two graphs share an isomorphism, developed by Tommi Junttila and Petteri Kaski Engineering an efficient canonical labeling tool for large and sparse graphs. The work provided here is theirs.

Its sources (C++) are provided on the project website under LGPL licence, and for convenience, a Python wrapper have also been published.

Unfortunately the bindings require Python 2.x, which is quite rare as a default environment by now.

We provide here a Dockerfile of this wrapper, to get you straight to a working installation, with a few little modifications of the original :

add installation of PyBliss module, so you don't have to manually add path, import PyBliss will work everywhere
add a parser function from_dimacs
add extra test script of canonical traces

Usage

Prerequisites

A working docker engine. See https://docs.docker.com/install/, or

curl -fsSL https://get.docker.com -o get-docker.sh
sh get-docker.sh

Get started

# build image
docker build -t pybliss .

# run the image in an interactive container
docker run -it --rm pybliss

From now you are inside a debian-based anaconda container, with PyBliss available in your favorite python interpreter, python or ipython.

Python usage

import PyBliss

G = PyBliss.from_dimacs('./data/ag/ag2-3-1')
H = PyBliss.from_dimacs('./data/ag/ag2-3-2')

print(str(G.relabel(G.canonical_labeling())))

str(H.relabel(H.canonical_labeling())) == str(G.relabel(G.canonical_labeling()))

See test_enumerate.py and test_iso.py for further test cases.