how to add nodes to an existing graph in networkx?

Question

Consider this code:

 G = nx.DiGraph()
 H=nx.path_graph(counter+1)
 G.add_nodes_from(H)
 labels = dict([ x for x in enumerate(idvalue) ])
 l =len(idvalue)
 labels[l]=userid
 for node in H:
     G.add_edge(node,l)
 pos=nx.spring_layout(G)
 nx.draw(G, pos=pos, node_color='g', node_size=20, with_labels=False)
 nx.draw_networkx_labels(G,pos,labels,font_size=16)
 plt.show()

Here, I'm extracting userid & idvalue from a URL using beautiful soup and these two values change as I iterate over a set of URLs. How to maintain one graph and add nodes as and when I extract content?

Please help. Thanks in advance.

Answer 1

you're already using add_edge() which adds the nodes as well if they don't exist.
if you want to add just a node with no edges attached you can use add_node()

Answer 2

Your problem is not that it's creating multiple graphs. Instead it's that you're plotting the same graph multiple times, and each time it looks different.

import networkx as nx
import matplotlib.pyplot as plt
G=nx.DiGraph()
G.add_edge(1,2)
labels = {1:1, 2:2}
for ctr in range(4):
    pos=nx.spring_layout(G)
    nx.draw(G, pos=pos, node_color='g', node_size=20, with_labels=True)
    nx.draw_networkx_labels(G,pos,labels,font_size=16)
    plt.savefig('fig'+str(ctr)+'.png')

You should see that it probably looks different each time it plots. That's because by default spring_layout has randomness built in. Since it's plotting in the same figure each time, it will look like you're getting lots of different graphs (note, how it handles the replotting if you do this in an interactive environment depends somewhat on that environment). You can fix this by adding plt.clf() before the plotting.

here's my first and last figures: