Graph rewriting and graph-based knowledge representation library. Documentation is available at http://dev.executableknowledge.org/ReGraph.
The ReGraph Python library is a generic framework for modelling graph-based systems. In this context models are viewed as graphs and graph transformations --- as a tool to describe both the system evolution and the model evolution read more about the approach.
ReGraph provides various utilities for graph rewriting which can be used for modelling the evolution of a system represented by a graph subject to rewriting. The rewriting functionality is based on the sesqui-pushout rewriting procedure [7]. In addition, the library enables a user to define a typing for models (graphs) that gives specifications for the structure of the models. This later functionality allows both to preserve the specified structure during rewriting and to propagate the changes to the specifications up to the models.
ReGraph contains a collection of utilities for rewriting graphs and hierarchies of graphs. It supports two backends: NetworkX graph objects and on Neo4j property graphs stored in a graph database.
In order to install the ReGraph library you can either use the latest version in PyPI:
pip install regraph
or install from the source by cloning the repository using SSH
git clone git@github.com:Kappa-Dev/ReGraph.git
or using HTTPS
https://github.com/Kappa-Dev/ReGraph.git
Install the library using Pythin setup tools:
python setup.py install
If you want to use the Neo4j-based backend of ReGraph, you need to install the Neo4j database (see installation instructions).
Moreover, ReGraph uses the APOC Neo4j plugin, currently not included in the community edition. To install the plugin see the instructions.
ReGraph uses Neo4j Bolt Driver for Python, therefore, having set up your database, you need to provide to ReGraph's API the address of the bolt server (for example, bolt://127.0.0.1:7687) and your credentials for connecting the database (i.e. user and password).
nosetests -v -s