nbgraph

Interactive graph exploration inside of a Jupyter notebook. Graphs are rendered with sigma.js.

Installation

Install is as simple as:

pip install nbgraph

Use

Graphs are represented with the NotebookGraph class. Although manual graph construction is possible with the add_node and add_edge methods, the anticipated use-case is that the user will create or load a graph with another library and use the nbgraph represenation only for the final display of the graph. With this in mind, nbgraph comes with loaders for working with graphs represented by igraph or networkx.

Quick-start Examples

See the examples directory for the examples as Jupyter notebooks.

igraph

from nbgraph.loaders.igraph_loader import load_graph
import igraph
G = igraph.Graph.Atlas(381)
load_graph(G)

networkx

from nbgraph.loaders.networkx_loader import load_graph
import networkx as nx
G=nx.dodecahedral_graph()
load_graph(G)

Notebooks with multiple graphs

If you use multiple graphs within the same notebook, you will see a warning about nbgraph being loaded multiple times. This is because the Python code can't reliably know what's loaded in the JavaScript, so it tries to load it every time a graph is displayed. This means sending the entire graph rendering code for each graph, which can explode the size of an .ipynb file.

To avoid this, use prepare_notebook:

from nbgraph.renderer import display_notebook, prepare_notebook
prepare_notebook()

Then, use display_notebook with an extra False argument to display the graph:

my_graph = load_graph(G)
display_notebook(load_graph(my_graph), False)