基于边宽度的matplotlib图例

我拥有以下网络:

我想根据边的宽度创建一个图例。如下(数字不准确):

我应该创建一个新的子图,自己绘制线条和宽度吗?使用这种方法,我发现它非常复杂,因为有一些边缘情况需要处理,还有一些残酷的假设,比如要绘制的线的数量(尽管我想假设有8~10+条不同宽度的线是可以的)。

import networkx as nx
import matplotlib.pyplot as plt

plt.subplot(1, 2, 1)
graph = nx.Graph()
graph.add_edge('a','b',weight=6)
graph.add_edge('a','c',weight=2)
graph.add_edge('c','d',weight=1)
graph.add_edge('c','e',weight=7)
graph.add_edge('c','f',weight=9)
graph.add_edge('a','d',weight=3)
edges = graph.edges()
edges_weight_list = [graph[u][v]['weight'] for u,v in edges]
n_nodes = graph.number_of_nodes()
pos = nx.spring_layout(graph)
nx.draw_networkx_edges(graph, pos, width = edges_weight_list)
nx.draw_networkx_labels(graph, pos)
mcp = nx.draw_networkx_nodes(graph, pos,
                             node_color=list(range(n_nodes)),
                             cmap='Blues')
limits = plt.axis('off')  # turn of axis


# width lines
plt.subplot(1, 2, 2)
edges_weight_list = sorted(edges_weight_list)
for i, current_weight in enumerate(edges_weight_list):
    x=[0, 1]
    y=[i, i]
    plt.plot(x,y, linewidth=current_weight, color='black')

plt.colorbar(mcp)
limits = plt.axis('off')  # turn of axis
plt.show()

该图:

(显然,在一个真实的例子中,我不会迭代所有的行)。

我使用的是python3.6、networkx2.2、matplotlib2.2.2。