Assign a color to each path in NetworkX

问题

I have cars,cities and routes. Every city is a node. Every route is a path generated by a car.

Different cars will have different path, sometimes paths could be intersected (which means differents cars have found the same city in they path), sometimes not.

I would rappresent a graph with all the cities and all the different path and plot the graph with plotly. Example:

List of cities: CityA -CityB -CityD -CityZ -CityK
List of cars: Car1, Car2

Routes:
Car1 will have a path through   cityA - cityB - cityD  this path will be colored in red
Car2 will have a path though    cityZ - cityA - cityK  this path will be colored in blue

Using networkx.classes.function.add_path I can't achive this because I will not preserve the information about different cars, there will be only the list of connected node:

As in the previous example add_path, G.edges(): [(CityA-CityB),(CityB-CityD),(CityZ-CityA),(CityA-CityK)]

I am not sure if what I am looking for could be achived with networkx.

A solution to plot is just passing the list to plotly but doing so I will not even use NetworkX and the next steps is to analize the graph.

回答1:

You can set node and edge attributes in NetworkX, which you can then use for instance to customize certain aspects of the plot. In this case, you could set a color attribute to the edges of the graph, and use this attribute to set the edge_color in nx.draw. Here's how you could do this with the example paths you've shared:

import networkx as nx
from matplotlib import pyplot as plt

path_car1 = ['cityA','cityB','cityD']
path_car2 = ['cityZ','cityA','cityK']

paths = [path_car1, path_car2]
colors = ['Red','Blue']

Now create a directed graph, and iterate over the lists of paths, and assigned colors to add them as edges, an corresponding attributes:

G = nx.DiGraph()
for path, color in zip(paths, colors):
    for edge in zip(path[:-1], path[1:]):
        G.add_edge(*edge, color=color)

You can get the values of a given attribute for all edges with:

edge_colors = nx.get_edge_attributes(G, 'color')

Now when plotting you can set the edge colors through the edge_color argument:

plt.figure(figsize=(10,7))
pos = nx.spring_layout(G, scale=20)
nx.draw(G, pos, 
        node_color='black',
        with_labels=True, 
        node_size=1200,
        edgelist=G.edges(),
        edge_color=edge_colors.values(),
        arrowsize=15,
        font_color='white',
        width=3,
        alpha=0.9)

来源：https://stackoverflow.com/questions/61599058/assign-a-color-to-each-path-in-networkx

标签

python

graph

plotly

data-science

networkx