cookbook.rst

Simplest use case

Given a networkx.Graph object, you can launch netwulf like so:

import networkx as nx
import netwulf as nw

G = nx.barabasi_albert_graph(100, 2)

wulf.visualize(G)  # <-- THIS IS IT

Alternatively, netwulf.visualize can accept a node-link dictionary object formatted like this.

Node and link attributes

Netwulf recognizes node attributes 'group' and 'size', and link attribute 'weight'. Users can create a networkx.Graph object with node and link data:

list(G.nodes(data=True))[:3]
# [(0, {'group': 0, 'size': 0.20982489558943607}),
#  (1, {'group': 0, 'size': 0.7118952904573288}),
#  (2, {'group': 0, 'size': 0.8785902846905586})]

list(G.edges(data=True))[:3]
# [(0, 5, {'weight': 0.8917083938103719}),
#  (0, 9, {'weight': 0.29583879684946757}),
#  (0, 12, {'weight': 0.36847140599448236})]

Example:

import numpy as np
import networkx as nx
import netwulf as nw

# Create a network
G = nx.random_partition_graph([10, 10, 10], .25, .01)

# Change 'block' node attribute to 'group'
for k, v in G.nodes(data=True):
    v['group'] = v['block']; del v['block']

# Or detect communities and encode them in 'group' attribute
# import community
# bb = community.best_partition(G)
# nx.set_node_attributes(G, bb, 'group')

# Set node 'size' attributes
for n, data in G.nodes(data=True):
    data['size'] = np.random.random()

# Set link 'weight' attributes
for n1, n2, data in G.edges(data=True):
    data['weight'] = np.random.random()

nw.visualize(G)

Note: If 'group' is not a color (like "red" or "#4fba21") the group colors are assigned randomly.

Initial node positions

A network can be launched with initial node positions. If netwulf sees node-attributes 'x' and 'y' like:

list(G.nodes(data=True))[:3]
# [(0, {'x': 600, 'y': 400}),
#  (1, {'x': 550, 'y': 450}),
#  (2, {'x': 500, 'y': 500})]

it freezes the nodes in these positions at launch. Nodes can be moved around in their frozen states. Positions are relaxed upon untoggling "Freeze", toggling "Wiggle" or changing any of the physics parameters.

Save as PDF

import networkx as nx
import netwulf as nw
import matplotlib.pyplot as plt

G = nx.barabasi_albert_graph(100, 2)

network, config = nw.visualize(G, plot_in_cell_below=False)

fig, ax = nw.draw_netwulf(network, figsize=(10,10))
plt.savefig("myfigure.pdf")

Labels and node positions

import networkx as nx
import netwulf as nw
import matplotlib.pyplot as plt

G = nx.Graph()
G.add_nodes_from([0,1,2,'a','b','c'])
G.add_edges_from([(0,1),('a','b')])

network, config = nw.visualize(G,config={'zoom':3})

# draw links only at first
fig, ax = nw.draw_netwulf(network,draw_nodes=False)

# get positions of two unconnected nodes to draw a link anyway
v0 = nw.node_pos(network, 'c')
v1 = nw.node_pos(network, 2)
ax.plot([v0[0],v1[0]],[v0[1],v1[1]],c='#d95f02')

# draw nodes now
nw.draw_netwulf(network,fig,ax,draw_links=False)

# add labels to a node and an edge
nw.add_node_label(ax,network,'c')
nw.add_edge_label(ax,network,('a','b'))