-
Notifications
You must be signed in to change notification settings - Fork 19
/
data.py
169 lines (124 loc) · 6.81 KB
/
data.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
import scipy.io as sio
import glob
from utils import *
def task_generator(feature, l_list, ul_list, bs, device):
feature_l = []
label_l = []
feature_l_qry = []
label_l_qry = []
for i in range(len(feature)):
perm_l = list(range(len(l_list[i])))
random.shuffle(perm_l)
perm_ul = list(range(len(ul_list[i])))
random.shuffle(perm_ul)
# generate support set
support_idx = np.array(l_list[i])[perm_l[:int(bs / 2)]].tolist()
support_idx += np.array(ul_list[i])[perm_ul[:int(bs / 2)]].tolist()
label_t = np.concatenate((np.ones(int(bs / 2)), np.zeros(int(bs / 2))))
feature_l.append(feature[i][support_idx].to(device))
label_l.append(torch.FloatTensor(label_t).to(device))
# generate query set
bs_qry = 2 * (len(l_list[i]) - int(bs / 2))
qry_idx = np.array(l_list[i])[perm_l[-int(bs_qry / 2):]].tolist()
qry_idx += np.array(ul_list[i])[perm_ul[-int(bs_qry / 2):]].tolist()
label_t_qry = np.concatenate((np.ones(int(bs_qry / 2)), np.zeros(int(bs_qry / 2))))
feature_l_qry.append(feature[i][label_t_qry].to(device))
label_l_qry.append(torch.FloatTensor(label_t_qry).to(device))
return feature_l, label_l, feature_l_qry, label_l_qry
def test_task_generator(feature, l_list, ul_list, bs, label, test_idx, device):
feature_l = []
label_l = []
for q in range(3):
perm_l = list(range(len(l_list)))
random.shuffle(perm_l)
perm_ul = list(range(len(ul_list)))
random.shuffle(perm_ul)
# generate support set
support_idx = np.array(l_list)[perm_l[:int(bs / 2)]].tolist()
support_idx += np.array(ul_list)[perm_ul[:int(bs / 2)]].tolist()
label_t = np.concatenate((np.ones(int(bs / 2)), np.zeros(int(bs / 2))))
feature_l.append(feature[support_idx].to(device))
label_l.append(torch.FloatTensor(label_t).to(device))
return feature_l, label_l, feature[test_idx].to(
device), torch.FloatTensor(label[test_idx]).to(device)
def test_task_generator_backup(feature, l_list, ul_list, bs, label, test_idx, device):
perm_l = list(range(len(l_list)))
random.shuffle(perm_l)
perm_ul = list(range(len(ul_list)))
random.shuffle(perm_ul)
# generate support set
support_idx = np.array(l_list)[perm_l[:int(bs / 2)]].tolist()
support_idx += np.array(ul_list)[perm_ul[:int(bs / 2)]].tolist()
label_t = np.concatenate((np.ones(int(bs / 2)), np.zeros(int(bs / 2))))
return feature[support_idx].to(device), torch.FloatTensor(label_t).to(device), feature[test_idx].to(
device), torch.FloatTensor(label[test_idx]).to(device)
def load_yelp(file):
data = sio.loadmat(file)
network = data['Network'].astype(np.float)
labels = data['Label'].flatten()
attributes = data['Attributes'].astype(np.float)
return network, attributes, labels
def sp_matrix_to_torch_sparse_tensor(sparse_mx):
"""Convert a scipy sparse matrix to a torch sparse tensor."""
sparse_mx = sparse_mx.tocoo().astype(np.float32)
indices = torch.from_numpy(
np.vstack((sparse_mx.row, sparse_mx.col)).astype(np.int64))
values = torch.from_numpy(sparse_mx.data)
shape = torch.Size(sparse_mx.shape)
return torch.sparse.FloatTensor(indices, values, shape)
class DataProcessor:
def __init__(self, num_graph, degree, data_name):
self.num_graph = num_graph # number of auxiliary graph + 1 (target graph)
self.degree = degree
self.data_name = data_name
self.feature_l, self.label_l, self.adj_l = [], [], []
self.target, self.target_adj, self.target_feature, self.target_label= None, None, None, None
self.target_idx_train_ano_all, self.target_idx_train_normal_all, self.target_idx_val, self.target_idx_test = None, None, None, None
self.labeled_idx_l, self.unlabeled_idx_l = [], []
self.target_labeled_idx, self.target_unlabeled_idx = [], []
def data_loader(self):
l = glob.glob("graphs/{}/*.mat".format(self.data_name))
f_l = random.sample(l, self.num_graph)
random.shuffle(f_l)
for f in f_l[:-1]:
adj, feature, label = load_yelp(f)
adj = normalize_adjacency(adj)
adj = sp_matrix_to_torch_sparse_tensor(adj).float()
feature = torch.FloatTensor(feature.toarray())
feature = sgc_precompute(feature, adj, self.degree)
self.feature_l.append(feature)
self.label_l.append(label)
self.adj_l.append(adj)
# load the target graph
self.target = f_l[-1]
adj, feature, label = load_yelp(self.target)
adj = normalize_adjacency(adj)
self.target_adj = sp_matrix_to_torch_sparse_tensor(adj).float()
self.target_feature = torch.FloatTensor(feature.toarray())
self.target_feature = sgc_precompute(self.target_feature, self.target_adj, self.degree)
self.target_label = label
# split the target graph into train/valid/test with 4/2/4
idx_anomaly = np.random.permutation(np.nonzero(self.target_label == 1)[0])
idx_normal = np.random.permutation(np.nonzero(self.target_label == 0)[0])
split_ano = int(0.4 * len(idx_anomaly))
split_normal = int(0.4 * len(idx_normal))
self.target_idx_train_ano_all = idx_anomaly[:split_ano]
self.target_idx_train_normal_all = idx_normal[:split_normal]
self.target_idx_val = np.concatenate((idx_anomaly[split_ano:-split_ano], idx_normal[split_normal:-split_normal])).tolist()
self.target_idx_test = np.concatenate((idx_anomaly[-split_ano:], idx_normal[-split_normal:])).tolist()
print("data loading finished.")
def sample_anomaly(self, num_labeled_ano):
for i in range(self.num_graph - 1):
# sampling anomalies from auxiliary graphs
label_tmp = self.label_l[i]
idx_anomaly = np.random.permutation(np.nonzero(label_tmp == 1)[0])
idx_normal = np.random.permutation(np.nonzero(label_tmp == 0)[0])
self.labeled_idx_l.append(idx_anomaly[:num_labeled_ano].tolist())
self.unlabeled_idx_l.append(np.concatenate((idx_normal, idx_anomaly[num_labeled_ano:])).tolist())
self.target_idx_train_ano_all = np.random.permutation(self.target_idx_train_ano_all)
self.target_idx_train_normal_all = np.random.permutation(self.target_idx_train_normal_all)
if num_labeled_ano <= len(self.target_idx_train_ano_all):
self.target_labeled_idx = self.target_idx_train_ano_all[:num_labeled_ano].tolist()
self.target_unlabeled_idx = np.concatenate((self.target_idx_train_normal_all, self.target_idx_train_ano_all[num_labeled_ano:])).tolist()
return [self.feature_l, self.labeled_idx_l, self.unlabeled_idx_l], \
[self.target_feature, self.target_labeled_idx, self.target_unlabeled_idx]