-
Notifications
You must be signed in to change notification settings - Fork 0
/
main.py
535 lines (454 loc) · 16.8 KB
/
main.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
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
import tensorflow as tf
import numpy as np
import pickle as pkl
import scipy.sparse as sp
from scipy.sparse.linalg.eigen.arpack import eigsh
import sys
import tensorflow.keras
from sklearn.model_selection import train_test_split
import pandas as pd
from tensorflow.contrib import rnn
from sklearn.model_selection import train_test_split
from sklearn.metrics import f1_score, accuracy_score, recall_score, precision_score
import networkx as nx
import random
import matplotlib.pyplot as plt
import sklearn
import random
import pandas
from sklearn.model_selection import KFold
from sklearn.preprocessing import MinMaxScaler
from sklearn.svm import SVR
from sklearn import preprocessing
import scipy as sc
import os
import re
import gc
import itertools
import statistics
import pickle
import argparse
import argparse
import random
from numpy.random import seed
import os
from sklearn.model_selection import KFold
from sklearn.model_selection import StratifiedKFold
os.environ["CUDA_VISIBLE_DEVICES"]="-1"
import tensorflow as tf
from hierarchical import *
from periphery import *
from logisticreg import *
seed(123)
random.seed(123)
np.random.seed(123)
# In[2]:
def glrt_init(shape, name=None):
init_range = np.sqrt(6.0/(shape[0]+shape[1]))
initial = tf.random_uniform(shape, minval=-init_range, maxval=init_range, dtype=tf.float32)
return tf.Variable(initial, name=name)
def one(i,n):
a = np.zeros(n, 'uint8')
a[i] = 1
return a
################################ THIS FUNCTION (read_graphfile) IS ADAPTED FROM RexYing/HybridPool ############################################
def read_graphfile(dataname):
max_nodes=None
#read datasets
prefix='dataset_graph/'+dataname+'/'+dataname
data_list=[]
data={}
filename_graph_indic = prefix + '_graph_indicator.txt'
# index of graphs that a given node belongs to
graph_indic={}
with open(filename_graph_indic) as f:
i=1
for line in f:
line=line.strip("\n")
graph_indic[i]=int(line)
i+=1
filename_nodes=prefix + '_node_labels.txt'
node_labels=[]
try:
with open(filename_nodes) as f:
for line in f:
line=line.strip("\n")
node_labels+=[int(line) - 1]
num_unique_node_labels = max(node_labels) + 1
except IOError:
print('No node labels')
filename_node_attrs=prefix + '_node_attributes.txt'
node_attrs=[]
try:
with open(filename_node_attrs) as f:
for line in f:
line = line.strip("\s\n")
attrs = [float(attr) for attr in re.split("[,\s]+", line) if not attr == '']
node_attrs.append(np.array(attrs))
except IOError:
print('No node attributes')
label_has_zero = False
filename_graphs=prefix + '_graph_labels.txt'
graph_labels=[]
# assume that all graph labels appear in the dataset
#(set of labels don't have to be consecutive)
label_vals = []
with open(filename_graphs) as f:
for line in f:
line=line.strip("\n")
val = int(line)
if val not in label_vals:
label_vals.append(val)
graph_labels.append(val)
label_map_to_int = {val: i for i, val in enumerate(label_vals)}
graph_labels = np.array([label_map_to_int[l] for l in graph_labels])
filename_adj=prefix + '_A.txt'
adj_list={i:[] for i in range(1,len(graph_labels)+1)}
index_graph={i:[] for i in range(1,len(graph_labels)+1)}
num_edges = 0
with open(filename_adj) as f:
for line in f:
line=line.strip("\n").split(",")
e0,e1=(int(line[0].strip(" ")),int(line[1].strip(" ")))
adj_list[graph_indic[e0]].append((e0,e1))
index_graph[graph_indic[e0]]+=[e0,e1]
num_edges += 1
for k in index_graph.keys():
index_graph[k]=[u-1 for u in set(index_graph[k])]
graphs=[]
for i in range(1,1+len(adj_list)):
# indexed from 1 here
G=nx.from_edgelist(adj_list[i])
if max_nodes is not None and G.number_of_nodes() > max_nodes:
continue
# add features and labels
G.graph['label'] = graph_labels[i-1]
for u in G.nodes():
if len(node_labels) > 0:
node_label_one_hot = [0] * num_unique_node_labels
node_label = node_labels[u-1]
node_label_one_hot[node_label] = 1
G.node[u]['label'] = np.array(node_label_one_hot)
if len(node_attrs) > 0:
G.node[u]['feat'] = node_attrs[u-1]
if len(node_attrs) > 0:
G.graph['feat_dim'] = node_attrs[0].shape[0]
mapping={}
it=0
if float(nx.__version__)<2.0:
for n in G.nodes():
mapping[n]=it
it+=1
else:
for n in G.nodes:
mapping[n]=it
it+=1
graphs.append(nx.relabel_nodes(G, mapping))
max_num_nodes = max([G.number_of_nodes() for G in graphs])
if len(node_attrs)>0:
feat_dim = graphs[0].node[0]['feat'].shape[0]
lab1=[]
feat_dim1 = graphs[0].node[0]['label'].shape[0]
for G in graphs:
adj = np.array(nx.to_numpy_matrix(G))
num_nodes = adj.shape[0]
adj_padded = np.zeros((max_num_nodes,max_num_nodes))
adj_padded[:num_nodes, :num_nodes] = adj
label1=G.graph['label']
if len(node_attrs)>0:
f = np.zeros((max_num_nodes,feat_dim), dtype=float)
for i,u in enumerate(G.nodes()):
f[i,:] = G.node[u]['feat']
f1=np.identity(max_num_nodes)
f = np.concatenate((f, f1), axis=1)
rowsum = np.array(f.sum(1))
r_inv = np.power(rowsum, -1).flatten()
r_inv[np.isinf(r_inv)] = 0.
r_mat_inv = np.diag(r_inv)
f = r_mat_inv.dot(f)
else:
max_deg = 10
f = np.zeros((max_num_nodes,feat_dim1), dtype=float)
for i,u in enumerate(G.nodes()):
f[i,:] = G.node[u]['label']
degs = np.sum(np.array(adj), 1).astype(int)
degs[degs>max_deg] = max_deg
feat = np.zeros((len(degs), max_deg + 1))
feat[np.arange(len(degs)), degs] = 1
feat = np.pad(feat, ((0, max_num_nodes - G.number_of_nodes()), (0, 0)),
'constant', constant_values=0)
f = np.concatenate((feat, f), axis=1)
f1=np.identity(max_num_nodes)
rowsum = np.array(f.sum(1))
r_inv = np.power(rowsum, -1).flatten()
r_inv[np.isinf(r_inv)] = 0.
r_mat_inv = np.diag(r_inv)
f = r_mat_inv.dot(f)
lab1.append(label1)
label1=one(label1,len(label_vals))
data={}
data['feat']=f
data['adj']=adj_padded
data['label']=label1
data_list.append(data)
return data_list, len(label_vals),max_num_nodes,lab1
# In[3]:
def periphery_1(adj,labels,feat,arguments,nclasses,max_num_nodes):
####MAIN FUNCTION####
placeholders={}
final={}
train_adj=adj
train_label=labels
train_feat=feat
placeholders={'gcn_dim':[128,128,64,64],'gcn_encoder_layers':2,'feat_dim':feat[0].shape[-1],'learning_rate':0.001,'num_nodes':np.array(adj).shape[-1],'emb_dim':feat[0].shape[-1],'nclasses':np.array(labels).shape[-1]}
a1=0
pat=20
tf.reset_default_graph()
tf.set_random_seed(123)
D=gin_gae(placeholders)
# idx=np.arrange(len(adj))
######Change batch_size according to the dataset######
batch_size=len(train_adj)
num_batches=int(len(train_adj)/batch_size)
D.gae_architechture()
# #####to set GPU ##########
# # config = tf.ConfigProto(device_count = {'GPU': 4})
# # config.gpu_options.allow_growth=False
sess = tf.Session()#config=config)
sess.run(tf.global_variables_initializer())
sess.run(tf.local_variables_initializer())
asqmx=0.0
step = 0
embeddings=[]
cost_val=[]
for epoch in range(2000):
embeddings=[]
trainavrloss = 0
trainavracc = 0
i10=0
tr_step = 0
for j in range(num_batches):
feed = {}
adj1=np.reshape(train_adj[j*batch_size:j*batch_size+batch_size],[-1])
pos_weight = float(adj1.shape[0] - adj1.sum()) / adj1.sum()
norm = adj1.shape[0]/ float((adj1.shape[0]- adj1.sum()) * 2)
feed['input_features1'] = train_feat[j*batch_size:j*batch_size+batch_size]
feed['input_adj1'] = train_adj[j*batch_size:j*batch_size+batch_size]
feed['input_labels1']=train_label[j*batch_size:j*batch_size+batch_size]
feed['keep']=0.1
feed['pos']=pos_weight
feed['norm']=norm
feed['train1']=True
summ1,embd=D.runn(sess,feed,"train")
trainavrloss += summ1
tr_step += 1
cost_val.append(trainavrloss/tr_step)
print('epoch:',epoch,' Training: loss::', trainavrloss/tr_step)
if epoch > 30 and cost_val[-1] > np.mean(cost_val[-(30+1):-1]):
print("Early stopping...")
break
feed = {}
adj1=np.reshape(train_adj[j*batch_size:j*batch_size+batch_size],[-1])
pos_weight = float(adj1.shape[0] - adj1.sum()) / adj1.sum()
norm = adj1.shape[0]/ float((adj1.shape[0]- adj1.sum()) * 2)
feed['input_features1'] = train_feat
feed['input_adj1'] = train_adj
feed['input_labels1']=train_label
feed['keep']=0
feed['pos']=pos_weight
feed['norm']=norm
feed['train1']=False
s,embeddings=D.runn(sess,feed,"test")
print("Training Completed")
sess.close()
return embeddings
def hierarchical_2(adj,labels,feat,EF,arguments,nclasses,max_num_node):
train_adj=adj
train_label=labels
train_feat=feat
tf.reset_default_graph()
tf.set_random_seed(123)
np.random.seed(123)
batch_size=len(train_adj)
embeddings_last=EF
s=train_feat-embeddings_last
s=np.array(s)
print(s.shape[-1])
s=np.mean(s,axis=1,keepdims=True)
ran=4
ran1=4
num_pool=2
clusrstio=0.5
placeholders={'feat_dim1':s.shape[-1],'batch_size':batch_size,'ran':ran,'gcn_dim':128,'emb_dim':128,'feat_dim':feat[0].shape[1],'learning_rate':0.001,'num_nodes':max_num_node,'num_pool':num_pool,'outp_dim':128,'clusrstio':clusrstio,'nclasses':nclasses}
D=hierarchical_repr(placeholders)
num_batches=int(len(train_adj)/batch_size)
D.hierarchical_repr_arch()
sess = tf.Session()
sess.run(tf.global_variables_initializer())
sess.run(tf.local_variables_initializer())
asqmx=0.0
step = 0
node_embeddings=[]
rand1=[ii for ii in range(batch_size)]
cost_val=[]
for epoch in range(10000):
randd1=[]
for kl in range(ran):
idx = np.random.permutation(np.arange(0, batch_size)).tolist()
randd1.append(idx)
randd2=[]
for kl in range(ran1):
idx = np.random.permutation(np.arange(0, batch_size)).tolist()
randd2.append(idx)
trainavrloss = 0
trainavracc = 0
i10=0
batch_size=len(train_adj)
tr_step = 0
tr_size = len(train_adj)
num_batches=int(len(train_adj)/batch_size)
for j in range(num_batches):
feed = {}
feed['input_features1'] = train_feat[j*batch_size:j*batch_size+batch_size]
feed['input_adj1'] = train_adj[j*batch_size:j*batch_size+batch_size]
feed['s']=s[j*batch_size:j*batch_size+batch_size]
feed['input_labels1']=train_label[j*batch_size:j*batch_size+batch_size]
feed['keep']=0.2
feed['train1']=True
feed['listt']=randd1
feed['listt1']=randd2
s1,s2,embd,em1=D.runn(sess,feed,"train")
trainavrloss += s1+s2
tr_step += 1
print('epoch',epoch,'Training: loss::', trainavrloss/tr_step)#, 'accuracy::',trainavracc/tr_step)
cost_val.append(trainavrloss/tr_step)
if epoch > 1550 and cost_val[-1] > np.mean(cost_val[-(1550+1):-1]):
print("Early stopping...")
break
print("Training Completed")
feed = {}
feed['input_features1'] = train_feat
feed['input_adj1'] = train_adj
feed['s']=s
feed['input_labels1']=train_label
feed['keep']=0.
feed['train1']=False
feed['listt']=randd1
feed['listt1']=randd2
embd=D.runn(sess,feed,"test")
tr_step += 1
node_embeddings=embd
sess.close()
feats=np.array(node_embeddings[-1])
print(feats.shape)
feats=np.reshape(feats,[train_feat.shape[0],feats.shape[-1]])
print(feats.shape)
l1=[np.where(r==1)[0][0] for r in train_label]
e=feats
return feats
def logistic_3(adj,labels,feat,lab1,feats,arguments,nclasses,max_num_node):
train_adj=adj
train_label=labels
train_feat=feat
lab1=np.array(lab1)
######Change batch_size according to the dataset######
pat=10
print(feats.shape)
kf=StratifiedKFold(n_splits=10,shuffle=True,random_state=0) #KFold(n_splits=10)
ep=[[] for ir in range(10)]
it=0
for train_index, test_index in kf.split(train_adj,lab1):
train_label,test_label=labels[train_index],labels[test_index]
train_feat,test_feat=feats[train_index],feats[test_index]
tf.reset_default_graph()
batch_size=len(train_feat)
tf.set_random_seed(123)
nclasses=train_label[0].shape[-1]
max_no_node=train_feat.shape[1]
placeholders={'batch_size':batch_size,'emb_dim':labels[0].shape[-1],'feat_dim':feats.shape[1],'learning_rate':0.05,'num_nodes':max_num_node,'nclasses':nclasses}
D=logistic(placeholders)
num_batches=int(len(train_feat)/batch_size)
D.logistic_build()
sess = tf.Session()
sess.run(tf.global_variables_initializer())
sess.run(tf.local_variables_initializer())
vlss_mn = np.inf
vacc_mx = 0.0
asqmx=0.0
step = 0
# it=0
for epoch in range(5000):
trainavrloss = 0
trainavracc = 0
vallossavr = 0
valaccravr = 0
i10=0
batch_size=len(train_feat)
tr_step = 0
tr_size = len(train_feat)
for j in range(num_batches):
feed = {}
feed['input_features1'] = train_feat[j*batch_size:j*batch_size+batch_size]
feed['input_labels1']=train_label[j*batch_size:j*batch_size+batch_size]
feed['keep']=0.3
feed['train1']=True
summ1,a1=D.runn(sess,feed,"train")
trainavrloss += summ1
trainavracc += a1
tr_step += 1
feed = {}
i10=0
batch_size=len(test_feat)
feed['input_features1'] = test_feat
feed['input_labels1']=test_label
feed['keep']=0
feed['train1']=False
summ,a=D.runn(sess,feed,"val")
ep[it].append(a*100)
it+=1
ep1=np.mean(ep,axis=0)
ep11=ep1.tolist()
epi=ep11.index(max(ep11))
print(epi,max(ep11))
def argument_parser():
parser = argparse.ArgumentParser(description="GraPHmax for graph classification")
parser.add_argument("-dt", "--dataset", type=str, help="name of the dataset", default="MUTAG")
parser.add_argument("-ll", "--lrl",type=float, default=0.05, help="learning rate logistic")
# parser.add_argument("-ss", "--sub_samp", type=int, default=12, help="number of subgraphs to be sampled")
parser.add_argument("-lp", "--lrp", type=float, default=0.001, help="learning rate periphery")
parser.add_argument("-lh", "--lrh",type=float, default=0.001, help="learning rate hierarchical")
parser.add_argument("-pr", "--pool_rt", type=float, default=0.1, help="pooling ratio")
parser.add_argument("-pl", "--pool_lay", type=int, default=2, help="number of pooling layers")
parser.add_argument("-ngpr", "--negpr", type=int, default=4, help="number of negative samples periphery")
parser.add_argument("-nghr", "--neghr", type=int, default=4, help="number of negative samples hierarchical")
parser.add_argument("-ed", "--embd_dim", type=int, default=128, help="embedding dimension")
parser.add_argument("-dr", "--dropout", type=float, default=0, help="dropout rate")
arguments = parser.parse_args()
return arguments
def read_dataset(dataset):
################### read in graphs ########################
datasets,n_classes,max_num_node,lab1=read_graphfile(dataset)
datasets=np.array(datasets)
return datasets,n_classes,max_num_node,lab1
def main():
arguments = argument_parser()
dataset,nclasses,max_num_nodes,lab1=read_dataset(arguments.dataset)
#################SEPERATE EACH COMPONENT###################
adj=[]
labels=[]
feat=[]
subgraphs1=[]
for i in range(len(dataset)):
adj.append(dataset[i]['adj'])
labels.append(dataset[i]['label'])
feat.append(dataset[i]['feat'])
print(len(adj),len(labels),len(feat))
adj=np.array(adj)
labels=np.array(labels)
feat=np.array(feat)
EF=periphery_1(adj,labels,feat,arguments,nclasses,max_num_nodes)
representation=hierarchical_2(adj,labels,feat,EF,arguments,nclasses,max_num_nodes)
logistic_3(adj,labels,feat,lab1,representation,arguments,nclasses,max_num_nodes)
#Main Function
if __name__ == "__main__":
main()