How to add a key to a graph drawing?

user_6160119 0
Viewed 397

I have created a graph using networkx. When I draw this graph the node labels do not fit (node names can be pretty long), so I have replaced them by numeric indices. The problem with this workaround is that the resulting drawing is harder to interpret as those numbers are meaningless. This is my code:

import networkx as nx
import matplotlib.pyplot as plt

G = nx.Graph()
G.add_nodes_from(['First', 'Second', 'Third'])
G.add_edges_from([('First', 'Second'), ('First', 'Third'), ('Second', 'Third')])
name2num = {name: num + 1 for num, name in enumerate(list(G.nodes))}
H = nx.relabel_nodes(G, mapping=name2num, copy=True)

fig, (ax0, ax1) = plt.subplots(1, 2)
nx.draw(G, ax=ax0, with_labels=True)
nx.draw(H, ax=ax1, with_labels=True)

graphs
My question is: how could I add a key to the drawing on the right?
This is the information I'd like to display:

1 - First
2 - Second
3 - Third

I tried to add a legend, to no avail:

ax0.legend(list(name2num.values()), list(name2num.keys()))

PS: I found a number of threads related to my problem. The proposed solutions consist in encoding the node labels through different colors. I'd rather prefer to encode the node labels through numbers.

networkx legend matplotlib graph-theory python
2 Answers

Instead of using a ax.legend you could place a text box:

import networkx as nx
import matplotlib.pyplot as plt

G = nx.Graph()
G.add_nodes_from(["First", "Second", "Third"])
G.add_edges_from([("First", "Second"), ("First", "Third"), ("Second", "Third")])
name2num = {name: num + 1 for num, name in enumerate(list(G.nodes))}
H = nx.relabel_nodes(G, mapping=name2num, copy=True)

fig, (ax0, ax1) = plt.subplots(1, 2)
nx.draw(G, ax=ax0, with_labels=True)
nx.draw(H, ax=ax1, with_labels=True)


legend_text = "\n".join(f"{v} - {k}" for k, v in name2num.items())
props = dict(boxstyle="round", facecolor="w", alpha=0.5)
ax1.text(
    1.15,
    0.95,
    legend_text,
    transform=ax1.transAxes,
    fontsize=14,
    verticalalignment="top",
    bbox=props,
)

Network figure with text box legend

user_32797160

You could use nx.draw_networkx_labels and use the defined dictionary name2num to map the labels in the graph on the second axis. Note that you don't need to create a new graph with renamed labels, just plot the same, disabling with_labels and instead using the dictionary:

name2num = {name: f'{num}-{name}' for num, name in enumerate(list(G.nodes), 1)}

With this you'd get:

fig, (ax0, ax1) = plt.subplots(1, 2)

posG_renamed = nx.spring_layout(G)

nx.draw(G, ax=ax0, with_labels=True,
       node_size = 2000, node_color='lightblue')

nx.draw(G, pos=posG_renamed, ax=ax1, 
        node_size = 2000, node_color='lightgreen')

nx.draw_networkx_labels(G, posG_renamed, ax=ax1,
                        labels=name2num,
                        font_size=12)

enter image description here

user_96986840
Related