-
Notifications
You must be signed in to change notification settings - Fork 15
/
attTransGAE.py
174 lines (139 loc) · 6.36 KB
/
attTransGAE.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
from __future__ import division
from __future__ import print_function
import time
import tensorflow as tf
from sklearn.metrics import average_precision_score, roc_auc_score,precision_recall_fscore_support
from gcn.utils import *
from gcn.transductiveModels import TransductiveGraphEncoder
from attentionModel import AttTransGAE
# Set random seed
seed = 123
np.random.seed(seed)
tf.set_random_seed(seed)
# Settings
flags = tf.app.flags
FLAGS = flags.FLAGS
flags.DEFINE_string('dataset', 'ddi', 'Dataset string.')
flags.DEFINE_string('model', 'attTransGAE', 'Model string.') # 'attTransGAE', 'transGAE'
flags.DEFINE_float('learning_rate', 0.01, 'Initial learning rate.')
flags.DEFINE_integer('epochs', 200, 'Number of epochs to train.')
flags.DEFINE_integer('hidden0', 128, 'Number of units in hidden layer 0.')
flags.DEFINE_integer('hidden1', 64, 'Number of units in hidden layer 1.')
flags.DEFINE_integer('hidden2', 32, 'Number of units in hidden layer 2.') #dimension of the node embeddings
flags.DEFINE_float('dropout', 0.2, 'Dropout rate (1 - keep probability).')
flags.DEFINE_float('weight_decay', 5e-4, 'Weight for L2 loss on embedding matrix.')
flags.DEFINE_integer('early_stopping', 20, 'Tolerance for early stopping (# of epochs).')
flags.DEFINE_integer('max_degree', 3, 'Maximum Chebyshev polynomial degree.')
def ddi_load_data_GAE(dataset_str):
"""Load data."""
names = ['allx', 'ally', 'graph',"adjmat", "trainMask", "valMask", "testMask"]
objects = []
for i in range(len(names)):
with open("DDIdata/ind.{}.{}".format(dataset_str, names[i]), 'rb') as f:
if sys.version_info > (3, 0):
objects.append(pkl.load(f, encoding='latin1'))
else:
objects.append(pkl.load(f))
x, y, graphs, adjmats, train_mask, val_mask, test_mask = tuple(objects)
tmpx = np.multiply(y,train_mask+train_mask.T)
#features = sp.coo_matrix(tmpx).tolil()
features = tmpx
adjs = []
for adjmat in adjmats:
adjs.append(adjmat)
return adjs, features, tmpx, y, train_mask, val_mask, test_mask
# Load data
adjs, densefeatures, x, y, train_mask, val_mask, test_mask = ddi_load_data_GAE(FLAGS.dataset)
# print(sum(y[np.where(test_mask>0)]))
# Some preprocessing
#features = preprocess_features(densefeatures)
#features = sparse_to_tuple(densefeatures)
features = densefeatures
if FLAGS.model == 'transGAE':
support = []
for adj in adjs:
support.append(preprocess_adj_dense(adj))
num_supports = len(support)
model_func = TransductiveGraphEncoder
elif FLAGS.model == 'attTransGAE':
support = []
for adj in adjs:
support.append(preprocess_adj_dense(adj))
num_supports = len(support)
model_func = AttTransGAE
else:
raise ValueError('Invalid argument for model: ' + str(FLAGS.model))
# Define placeholders
placeholders = {
'support': [tf.placeholder(tf.float32) for _ in range(num_supports)],
'features': tf.placeholder(tf.float32, shape=(None, features.shape[1])),
'x_input': tf.placeholder(tf.float32, shape=(None, x.shape[1])),
'labels': tf.placeholder(tf.float32, shape=(None, y.shape[1])),
'labels_mask': tf.placeholder(tf.int32),
'test_mask': tf.placeholder(tf.int32),
'dropout': tf.placeholder_with_default(0., shape=()),
'num_features_nonzero': tf.placeholder(tf.int32) # helper variable for sparse dropout
}
input_dim = features.shape[1]
# Create model
model = model_func(placeholders, input_dim=input_dim, logging=True)
# Initialize session
sess = tf.Session()
#
# Define model evaluation function
def evaluate(x, features, support, labels, mask, test_mask, placeholders):
t_test = time.time()
feed_dict_val = construct_feed_dict(x, features, support, labels, mask, placeholders)
feed_dict_val.update({placeholders['test_mask']: test_mask})
outs_val = sess.run([model.loss, model.accuracy, model.predict()], feed_dict=feed_dict_val)
return outs_val[0], outs_val[1], outs_val[2], (time.time() - t_test)
# Init variables
sess.run(tf.global_variables_initializer())
cost_val = []
# Construct feed dictionary
feed_dict = construct_feed_dict(x, features, support, y, train_mask, placeholders)
feed_dict.update({placeholders['dropout']: FLAGS.dropout})
feed_dict.update({placeholders['test_mask']: test_mask})
# Train model
for epoch in range(FLAGS.epochs):
print(epoch)
t = time.time()
# Training step
outs = sess.run([model.opt_op, model.loss, model.accuracy, model.activations, model.attention, model.mixedADJ, model.adjs, model.attADJ], feed_dict=feed_dict)
mixedADJ = outs[5]
attentions = outs[4]
adjs_model = outs[6]
attADJ_model = outs[-1]
test_preds = outs[3][-1]
test_preds = test_preds + test_preds.T
testsubs = np.where(test_mask > 0)
roc = roc_auc_score(y[testsubs], test_preds[testsubs])
prauc = average_precision_score(y[testsubs], test_preds[testsubs])
print("Test set results:", roc, prauc)
#summary_writer.add_summary(outs[0],epoch)
# for var in tf.trainable_variables():
# sess.run(var)
# Validation
cost, acc, _, duration = evaluate(x, features, support, y, train_mask, val_mask, placeholders)
cost_val.append(cost)
# Print results
# print("Epoch:", '%04d' % (epoch + 1), "train_loss=", "{:.5f}".format(outs[1]),
# "train_acc=", "{:.5f}".format(outs[2]), "val_loss=", "{:.5f}".format(cost),
# "val_acc=", "{:.5f}".format(acc), "time=", "{:.5f}".format(time.time() - t))
if epoch > FLAGS.early_stopping and cost_val[-1] > np.mean(cost_val[-(FLAGS.early_stopping+1):-1]):
# print("Early stopping...")
break
# print("Optimization Finished!")
# Testing
test_cost, test_acc, test_preds, test_duration = evaluate(x, features, support, y, train_mask, test_mask, placeholders)
# test_preds = np.maximum(test_preds, test_preds.T)
test_preds = np.add(test_preds, test_preds.T)/2.0
testsubs = np.where(test_mask>0)
pkl.dump([testsubs,test_preds,y, mixedADJ, attentions],open("resOUTPUT.pkl","wb"))
roc = roc_auc_score(y[testsubs], test_preds[testsubs])
prauc = average_precision_score(y[testsubs], test_preds[testsubs])
# print("Test set results:", "cost=", "{:.5f}".format(test_cost),
# "accuracy=", "{:.5f}".format(test_acc), "time=", "{:.5f}".format(test_duration))
print("AttTransGAE", roc, prauc)
# prec, recall, f, _ = precision_recall_fscore_support(y[testsubs]>=0.5, test_preds[testsubs]>=0.5, labels=[1,0])
# print(prec,recall,f)