-
Notifications
You must be signed in to change notification settings - Fork 0
/
ReFex_embedding.py
297 lines (223 loc) · 9.96 KB
/
ReFex_embedding.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
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
from __future__ import division
import networkx as nx
import numpy as np
import cPickle
import os, sys
import copy as copy
import create_neg_and_pos_net as negpos
dataset = "alpha"
def getMaxInOutEdges(G):
nodes = G.nodes()
max_in = 1
max_out = 1
max_ego = 1
max_ego_out = 1
for node_id in nodes:
max_in = max(max_in, G.degree(node_id))
max_out = max(max_out, G.out_degree(node_id))
ego_net = nx.ego_graph(G, node_id)
ego_edges = ego_net.number_of_edges()
max_ego = max(max_ego, ego_edges)
nbrs = nx.neighbors(G, node_id)
nbrs_total_edges = 0
for nbr in nbrs:
nbrs_total_edges += G.out_degree(nbr)
nbrs_total_edges += G.degree(nbr)
max_ego_out = max(max_ego_out, nbrs_total_edges - ego_edges)
return max_in, max_out, max_ego, max_ego_out
def loadRev2Score(folder, node_to_score_map):
files = os.listdir(folder)
file_count = len(files)
for file in files:
if dataset not in file:
file_count = file_count - 1
continue
f = open(folder + file, "r")
print file
for l in f:
ls = l.strip().split(",")
node_id = int(ls[0][1:])
# node_id = ls[0]
node_to_score_map[node_id] = float(ls[1])
# if node_id in node_to_score_map:
# node_to_score_map[node_id].append(float(ls[1]))
# else:
# node_to_score_map[node_id] = [0]
f.close()
return node_to_score_map, file_count
def getRev2FairnessScore(G):
"""
load the rev2 output, get the fairness scores
:return: a map of {node_id, (fairness media, fairness score)}
"""
node_to_fairness_map = {}
node_to_goodness_map = {}
nodes = G.nodes()
for node_id in nodes:
node_to_fairness_map[node_id] = 0.0
node_to_goodness_map[node_id] = 0.0
node_to_fairness_map, fairness_file_count = loadRev2Score("./rev2/results/fairness/", node_to_fairness_map)
node_to_goodness_map, goodness_file_count = loadRev2Score("./rev2/results/goodness/", node_to_goodness_map)
return (node_to_fairness_map, node_to_goodness_map, fairness_file_count, goodness_file_count)
def getEgoNetEdges(G, node_id, max_ego_out, max_ego):
"""
Get EgoNet edge number
:param G:
:param node_id:
:return: count of edge number
"""
ego_net = nx.ego_graph(G, node_id)
ego_edges = ego_net.number_of_edges()
nbrs = nx.neighbors(G, node_id)
nbrs_total_edges = 0
for nbr in nbrs:
nbrs_total_edges += G.out_degree(nbr)
nbrs_total_edges += G.degree(nbr)
ego_out_going_edges = nbrs_total_edges - ego_edges
return ego_edges / max_ego, ego_out_going_edges / max_ego_out
def removeProductNodes(G):
nodes = G.nodes()
for node_id in nodes:
if node_id not in rev2_fairness_score_map:
G.remove_node(node_id)
def findFeatures(G, node_id, max_in, max_out, max_ego, max_ego_out):
"""
Get 6 features for each node
:param G:
:param node_id:
:return: a tuple of 6 features
"""
# basic network features
# 1. out degree
out_degree = G.out_degree(node_id) / max_out
# 2. in degree
in_degree = G.degree(node_id) / max_in
# 3,4. ego net edge count, ego_out_going_edges
ego_net_edges_count, ego_out_going_edges = getEgoNetEdges(G, node_id, max_ego, max_ego_out)
# 5,6. Rev2 features (fairness media score, fairness)
# if node_id not in rev2_fairness_score_map:
# (fairness_media_score, fairness_score) = (1.0, 1.0)
# else:
# (fairness_media_score, fairness_score) = rev2_fairness_score_map[node_id]
features = [out_degree, in_degree, ego_net_edges_count, ego_out_going_edges]
return features
def initFeatures(G, G_pos, G_neg):
"""
Initialize featues for each node
:param G:
:return:
"""
# removeProductNodes(G)
# removeProductNodes(G_pos)
# removeProductNodes(G_neg)
max_in, max_out, max_ego, max_ego_out = getMaxInOutEdges(G)
max_in_pos, max_out_pos, max_ego_pos, max_ego_out_pos = getMaxInOutEdges(G_pos)
max_in_neg, max_out_neg, max_ego_neg, max_ego_out_neg = getMaxInOutEdges(G_neg)
nodes = G.nodes()
for node_id in nodes:
features = findFeatures(G, node_id, max_in, max_out, max_ego, max_ego_out)
features_pos = findFeatures(G_pos, node_id, max_in_pos, max_out_pos, max_ego_pos, max_ego_out_pos)
features_neg = findFeatures(G_neg, node_id, max_in_neg, max_out_neg, max_ego_neg, max_ego_out_neg)
print "features: ",len(features), features
print "features_neg: ",len(features_neg), features_neg
print "features_pos: ",len(features_pos), features_pos
features_rev2_faireness = rev2_fairness_score_map[node_id]
features_rev2_goodness = node_to_goodness_score_map[node_id]
G.node[node_id]["features"] = features
# features_pos.extend(features_neg)
G.node[node_id]["features_pos_neg"] = copy.copy(features_pos)
G.node[node_id]["features_pos_neg"].extend(copy.copy(features_neg))
if features_pos is None:
print("stop here")
# print (features_rev2_faireness, features_rev2_goodness)
# features_pos.append(features_rev2_faireness)
# features_pos.append(features_rev2_goodness)
print (len(features_pos),features_rev2_faireness, features_rev2_goodness)
print ("add features to rev2 ")
G.node[node_id]["features_pos_neg_rev2"]= features_pos
print len(G.node[node_id]["features_pos_neg_rev2"]),G.node[node_id]["features_pos_neg_rev2"]
G.node[node_id]["features_pos_neg_rev2"].extend(features_neg)
print len(G.node[node_id]["features_pos_neg_rev2"]),G.node[node_id]["features_pos_neg_rev2"]
G.node[node_id]["features_pos_neg_rev2"].append(features_rev2_faireness)
print len(G.node[node_id]["features_pos_neg_rev2"]),G.node[node_id]["features_pos_neg_rev2"]
G.node[node_id]["features_pos_neg_rev2"].append(features_rev2_goodness)
print len(G.node[node_id]["features_pos_neg_rev2"]),G.node[node_id]["features_pos_neg_rev2"]
print "features: ", len(features), features
print "features_neg: ",len(features_neg), features_neg
print "features_pos: ",len(features_pos), features_pos
print "features: ", len(G.node[node_id]["features"]), G.node[node_id]["features"]
print len(G.node[node_id]["features_pos_neg"]),G.node[node_id]["features_pos_neg"]
print len(G.node[node_id]["features_pos_neg_rev2"]),G.node[node_id]["features_pos_neg_rev2"]
def augmentFeatures(G, node_id, k, features, feature_count, feature_key):
# find its neighbors
# print (node_id, feature_key, k-1, feature_count)
nbr_ids = G.adj[node_id]
if len(features) > feature_count * 3 ** (k - 1):
return
sum_features = np.zeros(feature_count * 3 ** (k - 1))
degree = G.out_degree(node_id)
for nbr_id in nbr_ids:
nbr_features = G.node[nbr_id][feature_key]
if len(sum_features) != len(nbr_features[0:feature_count * 3 ** (k - 1)]):
print feature_key, feature_count, k, features
print sum_features
print nbr_features[0:feature_count * 3 ** (k - 1)]
sum_features = np.add(sum_features, nbr_features[0:feature_count * 3 ** (k - 1)])
# sum_features = np.add(sum_features, nbr_features)
# append sum and mean only if the target nodes features length <=3*k
new_features = features
if degree == 0:
new_features = np.append(features, np.zeros(feature_count * 3 ** (k - 1)*2))
else:
new_features = np.append(new_features, np.divide(sum_features, degree * 1.0))
new_features = np.append(new_features, sum_features)
# return new_features
G.node[node_id][feature_key] = new_features
def recursive(G, k):
"""
Iteration k times, features augumentation by mean and sum neighbors' features.
:param G:
:param k:
:return:
"""
nodes = G.nodes()
for node_id in nodes:
# print G.node[node_id]["features"]
features = G.node[node_id]["features"]
features_pos_neg = G.node[node_id]["features_pos_neg"]
features_pos_neg_rev2 = G.node[node_id]["features_pos_neg_rev2"]
G_feature_count = 4
G_neg_pos_count = 8
G_neg_pos_rev2_count = 2 + 8
augmentFeatures(G, node_id, k, features, G_feature_count, "features")
augmentFeatures(G, node_id, k, features_pos_neg, G_neg_pos_count, "features_pos_neg")
augmentFeatures(G, node_id, k, features_pos_neg_rev2, G_neg_pos_rev2_count,
"features_pos_neg_rev2")
# G.node[node_id]["features"] = new_features
# G.node[node_id]["features_pos_neg"] = new_features_pos_neg
# G.node[node_id]["features_pos_neg_rev2"] = new_features_pos_neg_rev2
def writeCsvEmbedding(feature_label):
fw = open("./results/%s_graph_embedding_vectors_%s.csv" % (dataset, feature_label), "w")
for node in list(G.nodes(data=feature_label)):
node_id = node[0]
features = node[1][feature_label]
fw.write("%s,%s\n" % (node_id, ",".join(str(e) for e in features)))
fw.close()
# main code start here
G, G_pos, G_neg = negpos.getPosNegNet(dataset)
rev2_fairness_score_map, node_to_goodness_score_map, fairness_file_count, goodness_file_count = getRev2FairnessScore(G)
print "Total G nodes=%d" % G.number_of_nodes()
print "Total G_pos nodes=%d" % G_pos.number_of_nodes()
print "Total G_neg nodes=%d" % G_neg.number_of_nodes()
# drawNxGraph(getEgoNet(G, node_id), "Ego_Net_sockpuppet_{0}_{1}".format(dataset, node_id))
initFeatures(G, G_pos, G_neg)
recursive(G, 1)
recursive(G, 2)
recursive(G, 3)
writeCsvEmbedding("features")
writeCsvEmbedding("features_pos_neg")
writeCsvEmbedding("features_pos_neg_rev2")
nx.write_gpickle(G, "./results/%s_graph_embedding_featured_graph.pkl" % (dataset))
print "\nUsers only"
print "Total nodes=%d" % G.number_of_nodes()
print "Total edges=%d" % G.number_of_edges()