-
Notifications
You must be signed in to change notification settings - Fork 12
/
weibo_preprocessing.py
132 lines (101 loc) · 4.07 KB
/
weibo_preprocessing.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
# -*- coding: utf-8 -*-
"""
Data from https://aminer.org/influencelocality
Extract network and diffusion cascades from Weibo
"""
import os
import time
import tarfile
from urllib.request import urlretrieve
def split_train_and_test(cascades_file):
"""
# Keeps the ids of the users that are actively retweeting
# Train time:(2011.10.29 -2012.9.28) and test time (2012.9.28 -2012.10.29)
"""
f = open(cascades_file)
ids = set()
train_cascades = []
test_cascades = []
counter = 0
for line in f:
date = line.split(" ")[1].split("-")
original_user_id = line.split(" ")[2]
retweets = f.next().replace(" \n","").split(" ")
#----- keep only the cascades and the nodes that are active in train (2011.10.29 -2012.9.28) and test (2012.9.28 -2012.10.29)
retweet_ids = ""
#------- last month at test
if int(date[0])==2012 and ((int(date[1])>=9 and int(date[2])>=28) or (int(date[1])==10 and int(date[2])<=29)):
ids.add(original_user_id)
cascade = ""
for i in range(0,len(retweets)-1,2):
ids.add(retweets[i])
retweet_ids = retweet_ids+" "+retweets[i]
cascade = cascade+";"+retweets[i]+" "+retweets[i+1]
#------- For each cascade keep also the original user and the relative day of recording (1-32)
date = str(int(date[2])+3)
op = line.split(" ")
op = op[2]+" "+op[1]
test_cascades.append(date+";" +op+cascade)
#------ The rest are used for training
elif (int(date[0])==2012 and int(date[1])<9 and int(date[2])<28) or (int(date[0])==2011 and int(date[1])>=10 and int(date[2])>=29):
ids.add(original_user_id)
cascade = ""
for i in range(0,len(retweets)-1,2):
ids.add(retweets[i])
retweet_ids = retweet_ids+" "+retweets[i]
cascade = cascade+";"+retweets[i]+" "+retweets[i+1]
if(int(date[1])==9):
date = str(int(date[2])-27)
else:
date = str(int(date[2])+3)
op = line.split(" ")
op = op[2]+" "+op[1]
train_cascades.append(date+";" +op+cascade)
counter+=1
if (counter % 10000==0):
print("------------"+str(counter))
f.close()
return train_cascades, test_cascades, ids
def download():
file_tmp = urlretreive("https://www.dropbox.com/s/r0kdgeh8eggqgd3/retweetWithoutContent.rar", filename=None)[0]
tar = tarfile.open(fileobj=file_tmp)
tar.extractall("total.csv")
file_tmp = urlretreive("https://www.dropbox.com/s/r0kdgeh8eggqgd3/graph_170w_1month.rar", filename=None)[0]
tar = tarfile.open(fileobj=file_tmp)
tar.extractall("graph_170w_1month.txt")
def weibo_preprocessing(path):
os.chdir(path)
download()
#------ Split the original retweet cascades
train_cascades, test_cascades, ids = split_train_and_test("total.txt")
#------ Store the cascades
print("Size of train:",len(train_cascades))
print("Size of test:",len(test_cascades))
with open("train_cascades.txt","w") as f:
for cascade in train_cascades:
f.write(cascade+"\n")
with open("test_cascades.txt","w") as f:
for cascade in test_cascades:
f.write(cascade+"\n")
#------ Store the active ids
f = open("active_users.txt","w")
for uid in ids:
f.write(uid+"\n")
f.close()
#------ Keep the subnetwork of the active users
g = open("weibo_network.txt","w")
f = open("graph_170w_1month.txt","r")
found = 0
idx=0
for line in f:
edge = line.replace("\n","").split(" ")
if edge[0] in ids and edge[1] in ids and edge[2]=='1':
found+=1
g.write(line)
idx+=1
if (idx%2000000==0):
print(idx)
print(found)
print("---------")
g.close()
f.close()