biofabric is a Python library implementing the BioFabric network visualization technique described in Longabaugh 2012, Combing the hairball with BioFabric: a new approach for visualization of large networks and at http://www.biofabric.org/
BioFabric is a new way to visualize networks in a simple, deterministic way, by laying out nodes and edges as rows and columns on a grid based on their degree. Such visualization allow for the quick identification of hubs, communities, and peculiar network topologies:
Complete graph, using a classical representation
Same graph, displayed using the BioFabric technique
Various examples of graphs displayed using BioFabric can be found at http://www.biofabric.org/gallery/index.html and in the examples/ subdirectory.
biofabric is provided as an easy_install and pip compliant package which can be installed as follows:
to install the most up to date (and potentially unstable) version, type either
easy_install https://github.com/ajmazurie/biofabric/archive/master.zip pip install https://github.com/ajmazurie/biofabric/archive/master.zip
to install a specific version, such as 0.1.0 (the latest stable version), type either
easy_install https://github.com/ajmazurie/biofabric/archive/0.1.0.zip pip install https://github.com/ajmazurie/biofabric/archive/0.1.0.zip
biofabric depends on two excellent libraries: NetworkX to manipulate networks, and PyX to produce an output in various formats (pdf, png, eps, jpg, etc.) following the BioFabric technique. easy_install will install these for you in case they are not already installed in your system.
Once biofabric installed, it can be used through the draw() function:
import biofabric # generate a complete graph of 10 nodes using the # networkx library; this is the example shown above import networkx g = networkx.generators.classic.complete_graph(10) # draw it, as a PDF document biofabric.draw(g, "complete_graph.pdf")
Documentation and additional examples can be found at https://github.com/ajmazurie/biofabric/wiki
biofabric is released under a MIT/X11 license.