-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathcompress_graph.py
170 lines (144 loc) · 6.69 KB
/
compress_graph.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
import pandas as pd
import numpy as np
import networkx as nx
import random
# create train graph
def read_graph(edges, nodes):
g = nx.Graph()
for i in range(nodes):
g.add_node(i)
for edge in edges:
origin_city = int(edge[0])
dest_city = int(edge[1])
if not g.has_edge(origin_city, dest_city):
g.add_edge(origin_city, dest_city, weight=0)
g.get_edge_data(origin_city, dest_city)['weight'] += edge[2]
return g
dataset = 'ucsocial'
dic = {'ploblogs': 1224, 'ucsocial': 1899, 'condmat': 16264, 'wiki': 90153}
nodes_nums = dic[dataset]
for ith in range(1):#here we just create one result document for example
edges = pd.read_csv('data/' + dataset + '/' + dataset + '_numpy' + str(ith) + '.csv')
edges.columns = ['source', 'target', 'weight']
g = read_graph(edges.values, nodes_nums)
print('graph '+str(g.number_of_nodes())+', '+str(g.number_of_edges()))
test_edges = pd.read_csv('data/' + dataset + '/' + dataset + 'sam' + str(ith) + '.csv').values
for testedge in test_edges:
# use "-" as flag for testedges
if not g.has_edge(int(testedge[0]), int(testedge[1])):
g.add_edge(int(testedge[0]), int(testedge[1]), weight=0)
g.get_edge_data(int(testedge[0]), int(testedge[1]))['weight'] += -testedge[2]
# degree table
deg_list = {}
for node in g.nodes:
deg = len(g.adj[node]) + (node in g.adj[node])
deg_list[node] = deg
edges = edges.iloc[:, :-1].values
# edges = edges[:, :-1].astype(np.int)
edges = pd.DataFrame(edges)
edges = edges[edges[0] != edges[1]].values
# candidate set
edges_set = edges.tolist()
edges_set = [tuple(x) for x in edges_set]
e = edges_set[0]
edges_set = set(edges_set)
i = 0
for layer in range(1):#here we just compress once for example
while len(edges_set) > 0:
min_deg_edge = 100000
for edge_item in edges_set:
edge_deg = deg_list[edge_item[0]] + deg_list[edge_item[1]] # edge_degree
if edge_deg < min_deg_edge:
min_deg_edge = edge_deg
e = edge_item
# print('e='+str(e)+', deg='+str(min_deg_edge))
e0, e1 = int(e[0]), int(e[1])
common_neighbors = list(nx.common_neighbors(g, e0, e1))
# check whether has "-" or not
has_test_edge = 0
for cn in common_neighbors:
if g.get_edge_data(e0, cn)['weight'] < 0 or g.get_edge_data(e1, cn)['weight'] < 0:
has_test_edge = 1
break
if has_test_edge == 1:
edges_set.remove(e)
continue
# no "-", then compress
Ni = list(nx.neighbors(g, e0))
Nj = list(nx.neighbors(g, e1))
if len(Ni) > len(Nj):
for n in Nj:
w = g.get_edge_data(e1, n)['weight']
if not g.has_edge(e0, n):
g.add_edge(e0, n, weight=0)
# 更新度表
deg_list[e0] = deg_list[e0] + 1
deg_list[n] = deg_list[n] + 1 # 不能抵消
g.get_edge_data(e0, n)['weight'] += w
deg_list[n] = deg_list[n] - 1 # 合并在一个循环内
g.remove_node(e1)
del deg_list[e1]
edges_set.remove(e)
oriset = edges_set
subset = set()
subset1 = set()
for ite in oriset:
if (ite[0] != e1) and (ite[1] != e1) and (ite[0] != e0) and (ite[1] != e0):
subset1.add(ite)
# oriset = oriset - subset
# edges_set = oriset
del oriset
edges_set = subset1
else: # 把Ni连接到j节点上,删除i节点
for n in Ni:
w = g.get_edge_data(e0, n)['weight']
if not g.has_edge(e1, n):
g.add_edge(e1, n, weight=0)
deg_list[e1] = deg_list[e1] + 1
deg_list[n] = deg_list[n] + 1 # 不能抵消
g.get_edge_data(e1, n)['weight'] += w
deg_list[n] = deg_list[n] - 1 # 合并在一个循环内
g.remove_node(e0)
del deg_list[e0]
edges_set.remove(e)
oriset = edges_set
subset = set()
subset1 = set()
for ite in oriset:
if (ite[0] != e1) and (ite[1] != e1) and (ite[0] != e0) and (ite[1] != e0):
subset1.add(ite)
# oriset = oriset - subset
# edges_set = oriset
del oriset
edges_set = subset1
i = i + 1
print('compress '+str(i))
print('g.info = '+str(g.number_of_nodes())+', '+str(g.number_of_edges()))
new_edge_table = nx.to_pandas_edgelist(g)
edges = new_edge_table[new_edge_table['weight'] > 0]
edges = edges.iloc[:, :-1]
edges = edges[edges['source'] != edges['target']].values
# next layer's candidate set
edges_set = edges.tolist()
edges_set = [tuple(x) for x in edges_set]
edges_set = set(edges_set)
print(str(ith) + str(layer) + ' complete!')
print('g.info = ' + str(g.number_of_nodes()) + ', ' + str(g.number_of_edges()))
new_edge_table = nx.to_pandas_edgelist(g)
aset = set(new_edge_table['source']) | set(new_edge_table['target'])
nodes = len(aset)
print('nodes = ', nodes)
node_dic = {}
for i, n in enumerate(aset):
node_dic[n] = i
new_edge_table['source'] = new_edge_table['source'].map(node_dic)
new_edge_table['target'] = new_edge_table['target'].map(node_dic)
# nodes = new_edge_table.values.max() + 1
nodes = pd.DataFrame([nodes])
nodes.to_csv('data/' + dataset + '/' + 'compress/' + dataset + '_allcompress_dis_deg_layer1' + str(ith), sep='\t', index=False, header=False)
new_train = new_edge_table[new_edge_table['weight'] > 0]
new_test = new_edge_table[new_edge_table['weight'] < 0]
new_test['weight'] = -new_test['weight']
new_train.to_csv('data/' + dataset + '/' + 'compress/' + dataset + '_allcompress_dis_deg_layer1' + str(ith), sep='\t', index=False, header=False, mode='a')
new_test.to_csv('data/' + dataset + '/' + 'compress/' + dataset + '_allcompress_sam_dis_deg_layer1' + str(ith)+'.csv', index=False)
print(str(ith) + ' complete!')