netwulf 0.0.10

Interactively visualize networks with Ulf Aslak's d3-tool from Python.

netwulf

This package provides an interface between

networkx

Graph objects and

Ulf Aslak’s interactive webapp

for simple and better network visualizations.

Install

pip install netwulf

Beware: netwulf only works with Python 3!

Example

Standard

Create a network and look at it

import networkx as nx
from netwulf import visualize

G = nx.barabasi_albert_graph(100,m=1)
visualize(G)

When the visualization was posted to Python in the Browser, the function actually returns a dictionary containing all style information of the stylized network. For instance,

import networkx as nx
from netwulf import visualize

G = nx.barabasi_albert_graph(5,m=1)
# the following only works if the user clicked "Post to Python" in the visualization.
network_properties = visualize(G)
print(network_properties)

A possible result from this code is

{   'height': 919,
    'linkAlpha': 0.5,
    'linkColor': '#7c7c7c',
    'links': [   {'link': [0, 1], 'width': 2},
                 {'link': [0, 2], 'width': 2},
                 {'link': [0, 3], 'width': 2},
                 {'link': [2, 4], 'width': 2}],
    'nodeStrokeColor': '#000000',
    'nodeStrokeWidth': 0.5,
    'nodes': [   {   'color': '#16a085',
                     'id': 0,
                     'pos': [442.2593240813363, 454.37557840980224],
                     'radius': 5.632111911473014},
                 {   'color': '#16a085',
                     'id': 1,
                     'pos': [481.8171665449027, 438.58151881303337],
                     'radius': 5.632111911473014},
                 {   'color': '#16a085',
                     'id': 2,
                     'pos': [446.98116529402097, 495.93002968604276],
                     'radius': 5.632111911473014},
                 {   'color': '#16a085',
                     'id': 3,
                     'pos': [437.9340114428489, 414.42238247014905],
                     'radius': 5.632111911473014},
                 {   'color': '#16a085',
                     'id': 4,
                     'pos': [488.5033653510982, 494.1926029578719],
                     'radius': 5.632111911473014}],
    'width': 919
}

Config

It’s possible to change the default settings which are

default_config = {
  'Apply heat (wiggle)': False,
  'Charge strength': -10,
  'Center gravity': 0.1,
  'Link distance': 10,
  'Link width': 2,
  'Link alpha': 0.5,
  'Node size': 10,
  'Node stroke size': 0.5,
  'Node size exponent': 0.5,
  'Link strength exponent': 0.1,
  'Link width exponent': 0.5,
  'Collision': False,
  'Node fill': '#16a085',
  'Node stroke': '#000000',
  'Link stroke': '#7c7c7c',
  'Label stroke': '#000000',
  'Show labels': False,
  'Zoom': 1.5,
  'Min. link weight %': 0,
  'Max. link weight %': 100
}

It’s done like so:

import networkx as nx
from netwulf import visualize

G = nx.barabasi_albert_graph(5000,m=1)
visualize(G,config={
        'Node size': 11,
        'Charge strength' : -0.8,
        'Link distance' : 10,
        'Link width' : 1,
        'Collision' : True,
    })

Attributes

Node attributes such as ‘group’ or ‘size’ that you define in your networkx.Graph are automatically visualized.

import networkx as nx
import community
from netwulf import visualize

G = nx.random_partition_graph([10,10,10],.25,.01)
bb = community.best_partition(G)  # dict of node-community pairs
nx.set_node_attributes(G, bb, 'group')

visualize(G)

Dev notes

The JS base code in /netwulf/js/ is a fork of Ulf Aslak’s interactive web app. If this repository is updated, change to /netwulf/js/, then do

git fetch upstream
git merge upstream/master
git commit -m "merged"
git push