dynamicalab.drawing.group_radial_layout

dynamicalab.drawing.group_radial_layout(G, groups, radius=1, margin=0.3141592653589793, phase=0.0, dx=0.0, dy=0.0)

This constructs a circular layout of nodes with groups. Each group is separated by a certain angle.

Parameters

G : Networkx Graph

A Networkx Graph or DiGraph.

groups : list of lists

A list of groups of nodes. Each element of groups must be a list of node_id.

radius : float : (default=1)

Radius of the circle. Could be useful if you plan to plot two graphs of different size on the same figure.

margin : float : (default=np.pi/10)

Angle space between groups. Increase to add space between groups.

phase : float : (default=0.0)

Initial phase of the circle. This mostly induce a rotation of the nodes on the circle.

dx : float : (default=0.0)

X translation of nodes.

dy : float : (default=0.0)

Y translation of nodes.

Returns

dict

Dictionary of positions that can be use directly with draw_networks. Each key is a node_id (from groups) and the value is a position [x,y].

{
    0 : [-0.5,0.1],
    1 : [-0.1, 0.5],
    ...
}

Raise

ValueError

Occurs if groups does not contain all nodes (low criterion).

Example

import networkx as nx
import dynamicalab.drawing as draw
import matplotlib.pyplot as plt
import seaborn as sns

# Generate network
N = 20
G = nx.erdos_renyi_graph(N,0.4)

sns.set(style="ticks")
fig = plt.figure(figsize=(2,2))
ax = plt.gca()

# Divide in three groups
groups = [range(6), range(6,14), range(14,N)]
pos = draw.group_radial_layout(G, groups)

node_colors = ['#e41a1c']*6+['#377eb8']*8+['#4daf4a']*(N-14)
draw.draw_networks(G, pos, ax, 
                  mu=0.00, 
                  arrow_scale=0.0, 
                  edge_color="black", 
                  node_color=node_colors, 
                  use_edge_weigth=False)