Hello, thanks for a repo. Actually, I don't find anything for creating interactive network visualization, except yours repo.

Add a setup.py to installing with pip or easy_install

Since we are limited by iPython in the overall canvas size, incorporating the D3 Pan and Zoom features would be awesome for making use of the limited space. It would allow us to move things around and get a full feel for how things work.

Great work by the way, being able to go from a python data structure directly to a graph, is awesome.

I'm trying to plot a simple graph with a node and a child. Sometimes it renders 3 nodes (2 of them connected). Other times it renders nothing. What am I doing wrong?

eg2 = EventfulGraph()
d32 = ForceDirectedGraph(eg2)

d32.width = 300
d32.height = 300

edges = set()
for name, rid, pid in [('root', 1, None),
                       ('c1', 2, 1)]:
    eg2.add_node(rid, label=name, 
                 fill='red', 
                 stroke="black", color='black',
                 font_size='8px', r=5)
    if pid: 
        edges = edges | {(rid, pid)}
for i1, i2 in edges:
    eg2.add_edge(int(i1), int(i2),
                 distance=50)
display(d32)

I don't remember the state of including Javascript rendered components on the nbviewer. Is that possible in general now? Is it possible with D3? We're thinking about converting all networkx examples to notebooks and I'm wondering if they could simply be rendered by nbviewer or if they need to be hosted for the graphics to work. Thanks!

Great job on the demos!

One improvement I think is doable before Thursday's talk: adding labels to the factorization demo.

I'm not sure what will work best, perhaps a combination of fixed labels for all nodes beyond a certain size, and on-hover ones for the tiny ones? Or offset labels if a digit only has very few prime factors?

Basically, play with ideas to help the audience understand what each node represents, and in this case it really needs to be a way of showing the numerical value.

thanks!

cc @ellisonbg

Thanks for this great code! We have been discussing it over at wrobstory/vincent#96

How would one render a simple, static graph? Ideally something like:

    g = networkx.erdos_renyi(100,.1)
    render(g)

Unfortunately, even using the "demo simple" notebook it appears I can't
even do:

    g = networkx.erdos_renyi(100,.1)
    G = EventfulGraph(g)
    d3 = ForceDirectedGraphWidget(G)
    floating_container.children = [d3]
    display(floating_container)

I understand I could do:

    G = EventfulGraph()

and build the network I want from there "manually", but that's painful and sometimes impractical.

Using just the same code, importing with

from d3networkx import ForceDirectedGraph, EventfulGraph

results in the error:

ModuleNotFoundError: No module named 'eventful_dict'

My guess is this is because this line should be a relative import.