How to add labels to nodes in a graph in networkx?

integer, networkx, partitioning, python

Solution

Since your `node_list` was composed of integers, your nodes got the string representation of those integers as labels. But your nodes can be any hashable object, not just integers. So the simplest thing to do would be to make your `node_list` the string representation of the items in `parlist`. (The items in `parlist` are lists, which are mutable and therefore not hashable. That's why we can't just use `parlist` as our `node_list`.)

There is also the function nx.relabel_nodes, which we could use instead, but I think just giving the nodes the right label in the first place is simpler.

import networkx as nx
import matplotlib.pyplot as plt


def partitions(num):
    final = [[num]]

    for i in range(1, num):
        a = num - i
        res = partitions(i)
        for ele in res:
            if ele[0] <= a:
                final.append([a] + ele)

    return final


def drawgraph(parlist):
    G = nx.Graph()
    length = len(parlist)
    print "length is %s\n" % length
    node_list = [str(item) for item in parlist]
    G.add_cycle(node_list)
    pos = nx.circular_layout(G)
    draw_lifted(G, pos)


def draw_lifted(G, pos=None, offset=0.07, fontsize=16):
    """Draw with lifted labels
    http://networkx.lanl.gov/examples/advanced/heavy_metal_umlaut.html
    """
    pos = nx.spring_layout(G) if pos is None else pos
    nx.draw(G, pos, font_size=fontsize, with_labels=False)
    for p in pos:  # raise text positions
        pos[p][1] += offset
    nx.draw_networkx_labels(G, pos)
    plt.show()

drawgraph(partitions(4))

yields

Problem

Original source