使用python绘制人人网好友关系图示例

代码依赖：networkx matplotlib

 #! /bin/env python
# -*- coding: utf-8 -*-

import urllib
import urllib2
import cookielib
import re
import cpickle as p
import networkx as nx
import matplotlib.pyplot as plt

__author__ = """reverland ([e-mail])"""

# control parameters,edit it here
## login
username = 'none'
password = 'none'
## control graphs, edit for better graphs as you need
label_flag = true # whether shows labels.note: configure your matplotlibrc for chinese characters.
remove_isolated = true # whether remove isolated nodes(less than iso_level connects)
different_size = true # nodes for different size, bigger means more shared friends
iso_level = 10
node_size = 40 # default node size

 
def login(username, password):
    """log in and return uid"""
    logpage = "http://www.renren.com/ajaxlogin/login"
    data = {'email': username, 'password': password}
    login_data = urllib.urlencode(data)
    cj = cookielib.cookiejar()
    opener = urllib2.build_opener(urllib2.httpcookieprocessor(cj))
    urllib2.install_opener(opener)
    res = opener.open(logpage, login_data)
    print "login now ..."
    html = res.read()
    #print html

    # get uid
    print "getting user id of you now"
    res = urllib2.urlopen("http://www.renren.com/home")
    html = res.read()
    # print html
    uid = re.search("'ruid':'(\\d+)'", html).group(1)
    # print uid
    print "login and got uid successfully"
    return uid

 
def getfriends(uid):
    """get the uid's friends and return the dict with uid as key,name as value."""
    print "get %s 's friend list" % str(uid)
    pagenum = 0
    dict1 = {}
    while true:
        targetpage = "http://friend.renren.com/getfriendlist.do?curpage=" + str(pagenum) + "&id=" + str(uid)
        res = urllib2.urlopen(targetpage)
        html = res.read()

        pattern = '<a href="http://www\\.renren\\.com/profile\\.do\\?id=(\\d+)"><img src="[\\s]*" alt="[\\s]*[\\s]\\((.*)\\)" />'

        m = re.findall(pattern, html)
        #print len(m)
        if len(m) == 0:
            break
        for i in range(0, len(m)):
            no = m[i][0]
            uname = m[i][1]
            #print uname, no
            dict1[no] = uname
        pagenum += 1
    print "got %s 's friends list successfully." % str(uid)
    return dict1

 
def getdict(uid):
    """cache dict of uid in the disk."""
    try:
        with open(str(uid) + '.txt', 'r') as f:
            dict_uid = p.load(f)
    except:
        with open(str(uid) + '.txt', 'w') as f:
            p.dump(getfriends(uid), f)
        dict_uid = getdict(uid)
    return dict_uid

 
def getrelations(uid1, uid2):
    """receive two user id, if they are friends, return 1, otherwise 0."""
    dict_uid1 = getdict(uid1)
    if uid2 in dict_uid1:
        return 1
    else:
        return 0

 
def getgraph(username, password):
    """get the graph object and return it.
you must specify a chinese font such as `simhei` in ~/.matplotlib/matplotlibrc"""
    uid = login(username, password)
    dict_root = getdict(uid) # get root tree

    g = nx.graph() # create a graph object
    for uid1, uname1 in dict_root.items():
        # encode chinese characters for matplotlib **important**
        # if you want to draw chinese labels,
        uname1 = unicode(uname1, 'utf8')
        g.add_node(uname1)
        for uid2, uname2 in dict_root.items():
            uname2 = unicode(uname2, 'utf8')
            # not necessary for networkx
            if uid2 == uid1:
                continue
            if getrelations(uid1, uid2):
                g.add_edge(uname1, uname2)

    return g

 
def draw_graph(username, password, filename='graph.txt', label_flag=true, remove_isolated=true, different_size=true, iso_level=10, node_size=40):
    """reading data from file and draw the graph.if not exists, create the file and re-scratch data from net"""
    print "generating graph..."
    try:
        with open(filename, 'r') as f:
            g = p.load(f)
    except:
        g = getgraph(username, password)
        with open(filename, 'w') as f:
            p.dump(g, f)
    #nx.draw(g)
    # judge whether remove the isolated point from graph
    if remove_isolated is true:
        h = nx.empty_graph()
        for sg in nx.connected_component_subgraphs(g):
            if sg.number_of_nodes() > iso_level:
                h = nx.union(sg, h)
        g = h
    # ajust graph for better presentation
    if different_size is true:
        l = nx.degree(g)
        g.dot_size = {}
        for k, v in l.items():
            g.dot_size[k] = v
        node_size = [g.dot_size[v] * 10 for v in g]
    pos = nx.spring_layout(g, iterations=50)
    nx.draw_networkx_edges(g, pos, alpha=0.2)
    nx.draw_networkx_nodes(g, pos, node_size=node_size, node_color='r', alpha=0.3)
    # judge whether shows label
    if label_flag is true:
        nx.draw_networkx_labels(g, pos, alpha=0.5)
    #nx.draw_graphviz(g)
    plt.show()

    return g

if __name__ == "__main__":
    g = draw_graph(username, password)