通过networkx并使用python在图中添加节点和边

Question

I need to create a graph of IP addresses and mark edges between them for which I wrote the code below. The list, content of the parseOutput(), looks something like this :

('172.16.254.128', '216.58.208.206')
('216.58.208.206', '172.16.254.128')
('172.16.254.128', '216.58.208.226')
('216.58.208.226', '172.16.254.128')
('172.16.254.128', '8.8.8.8')
('8.8.8.8', '172.16.254.128')
('172.16.254.128', '216.58.208.227')
('172.16.254.128', '216.58.208.227')
('216.58.208.227', '172.16.254.128')
('172.16.254.128', '216.58.208.227')
('172.16.254.128', '216.58.208.227')
...

when I run this code, I get the following error:

Traceback (most recent call last):
  File "test.py", line 40, in <module>
    g.add_nodes_from(nodeList)
  File "/usr/local/lib/python2.7/dist-packages/networkx/classes/graph.py", line 429, in add_nodes_from
    nn,ndict = n
ValueError: need more than 0 values to unpack

---

import logging, sys, struct, binascii, socket
from scapy.all import *
import networkx as nx
import matplotlib.pyplot as plt

pkts=rdpcap("pcapFile.pcap",20)

def parsePcap():
        IPList = [[] for i in range (20)]
        count = 0
        for pkt in pkts:
                #print pkt.summary()
                if pkt.haslayer(IP):
                        #proto = pkt.getLayer(IP).proto
                        x = pkt.getlayer(IP).src
                        y = pkt.getlayer(IP).dst
                        IPList[count]= (x,y)
                        #IPList[count].append(proto)
                        print IPList[count]
                        count += 1
        return IPList


parseOutput = parsePcap()
nodeList = parseOutput
edgeList = parseOutput

g = nx.Graph()

g.add_nodes_from(nodeList)
print g.number_of_nodes
g.add_edges_from(edgeList)

pos = nx.spring_layout(g,scale=1) #default to scale=1
nx.draw(g,pos)

I dont' even know what I'm doing wrong. The documentation doesn't tell much and all examples online seem to be written in the same syntaxt as my code.

Answer 1

The problem you are seeing is caused by

IPList = [[] for i in range (20)]

which leads to parsePcap() returning a list of sequences with an empty list or lists at the end when len(pkts) is less than 20:

parseOutput = [
('172.16.254.128', '216.58.208.206'),
...
('172.16.254.128', '216.58.208.227'),
[],  #<---------------- This is causing the problem 
]

When parseOutput is passed to g.add_nodes_from , you get the traceback error message:

File "/usr/local/lib/python2.7/dist-packages/networkx/classes/graph.py", line 429, in add_nodes_from
  nn,ndict = n
ValueError: need more than 0 values to unpack

In retrospect, if you think carefully about the error message you can see it is telling you that n has 0 values to unpack. That makes sense if n , the node, is an empty list:

In [136]: nn, ndict = []
ValueError: need more than 0 values to unpack

The empty list is coming from parseOutput .

---

Instead of pre-allocating a list of a fixed size:

IPList = [[] for i in range (20)]

the preferred way to do this in Python is to use the append method:

def parsePcap():
    IPList = []
    for pkt in pkts:
        if pkt.haslayer(IP):
            x = pkt.getlayer(IP).src
            y = pkt.getlayer(IP).dst
            IPList.append((x, y))
    return IPList

This is easier and more readable, since you do not need to fuss with index numbers and increment a counter variable. Moreover, it allows you to handle an arbitrary number of items in pkts without first having to know the length of pkts .

---

Another thing that needs to be fixed is that the nodeList typically is not the same as the edgeList .

If you are going to declare a nodeList , the nodeList should be an iterable of IP addresses, and the edgeList should be an iterable of tuples, like parseOutput :

nodeList = set([item for pair in parseOutput for item in pair])
print(nodeList)
# set(['216.58.208.206', '216.58.208.227', '172.16.254.128',
#  '216.58.208.226', '8.8.8.8'])

However, since all your nodes are also mentioned in edgeList, you can omit declaring the nodes and just use

edgeList = parseOutput
g.add_edges_from(edgeList)

g.add_edges_from will add the nodes implicitly.

---

import matplotlib.pyplot as plt
import networkx as nx

parseOutput = [
('172.16.254.128', '216.58.208.206'),
('216.58.208.206', '172.16.254.128'),
('172.16.254.128', '216.58.208.226'),
('216.58.208.226', '172.16.254.128'),
('172.16.254.128', '8.8.8.8'),
('8.8.8.8', '172.16.254.128'),
('172.16.254.128', '216.58.208.227'),
('172.16.254.128', '216.58.208.227'),
('216.58.208.227', '172.16.254.128'),
('172.16.254.128', '216.58.208.227'),
('172.16.254.128', '216.58.208.227'),]

g = nx.Graph()
edgeList = parseOutput
g.add_edges_from(edgeList)

pos = nx.spring_layout(g,scale=1) #default to scale=1
nx.draw(g,pos, with_labels=True)
plt.show()

yields

---