Graphine 0.0a1

What Is Graphine?

Graphine is a flexible, easy-to-use graph library for Python 3. It was developed with four key goals in mind:

1) No external dependencies- no graphics toolkits, no math libraries, just pure Python.

2) Python 3 readiness- if you’re looking at moving your application to Python 3, at least you won’t have to port your graph library too.

3) Flexibility- we aim to support many different graph use cases, from flow networks and planar graphs to tree walking and DAGs.

4) Ease of use- no graph library, however powerful, is useful if you can’t figure it out. We try to make that as painless as possible.

Who Should Use Graphine

1) Developers faced with a data set that naturally decomposes into a graph, but who are unwilling to reinvent graph theory to properly represent it. We make it easy to inspect graphs for special properties, add data and structural elements, and extend graphs for application-specific behavior.

2) Non-mathematicians who need to process complex data structures but aren’t familiar with the formal methods for doing so in the graph context. Graphine doesn’t require a strong mathematical background to use, and is flexible enough to permit graph problems to be represented in their most natural form.

3) Students and teachers looking for a clean, easy-to-understand example of how graph algorithms are supposed to behave.

4) Graph theorists looking for a tool to rapidly sketch out a complex problem without the hassle of manual memory management.

Features

Graphine is designed to be ready-to-use for most graph applications out of the box. It supports easy node and edge addition, modification, and removal, and provides attribute access to application-defined data. Formally, the graphs represented are bridge multigraphs, capable of representing both directed and undirected edges, parallel edges, identical nodes and edges, loops, and arbitrary node and edge labels.

It also provides a powerful set of graph operations, including a* traversals, BFS, DFS, shortest path, edge contraction, generators for random walks, and binary graph operations like union, intersection, and difference.

All of the above is present in the alpha package; experimental support for loading and storing graphs using GraphML and Dot is also included