-
Notifications
You must be signed in to change notification settings - Fork 2
/
datasets.py
195 lines (166 loc) · 7.7 KB
/
datasets.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
import networkx as nx
import numpy as np
import os
import pickle
import torch
import scipy.sparse as sp
from sklearn.model_selection import train_test_split
def process_features(features):
row_sum_diag = np.sum(features, axis=1)
row_sum_diag_inv = np.power(row_sum_diag, -1)
row_sum_diag_inv[np.isinf(row_sum_diag_inv)] = 0.
row_sum_inv = np.diag(row_sum_diag_inv)
return np.dot(row_sum_inv, features)
def sample_mask(idx, l):
mask = np.zeros(l)
mask[idx] = 1
return np.array(mask, dtype=np.bool)
def load_data1(dataset):
## get data
if dataset == 'cora' or dataset == 'citeseer' or dataset == 'pubmed':
data_path = 'data'
suffixs = ['x', 'y', 'allx', 'ally', 'tx', 'ty', 'graph']
objects = []
for suffix in suffixs:
file = os.path.join(data_path, 'ind.%s.%s'%(dataset, suffix))
objects.append(pickle.load(open(file, 'rb'), encoding='latin1'))
x, y, allx, ally, tx, ty, graph = objects
x, allx, tx = x.toarray(), allx.toarray(), tx.toarray()
# test indices
test_index_file = os.path.join(data_path, 'ind.%s.test.index'%dataset)
with open(test_index_file, 'r') as f:
lines = f.readlines()
indices = [int(line.strip()) for line in lines]
min_index, max_index = min(indices), max(indices)
# preprocess test indices and combine all data
tx_extend = np.zeros((max_index - min_index + 1, tx.shape[1]))
features = np.vstack([allx, tx_extend])
features[indices] = tx
ty_extend = np.zeros((max_index - min_index + 1, ty.shape[1]))
labels = np.vstack([ally, ty_extend])
labels[indices] = ty
labels1 = []
for i in range(len(labels)):
labels1.append(labels[i].argmax())
labels1 = np.array(labels1)
# get adjacency matrix
adj = nx.adjacency_matrix(nx.from_dict_of_lists(graph)).toarray()
# adj = torch.from_numpy(adj)
adj = np.array(adj)
idx_train = np.arange(0, len(y), 1)
idx_val = np.arange(len(y), len(y) + 500, 1)
idx_test = np.array(indices)
elif dataset == 'polblogs':
adj = np.zeros((1222, 1222))
with open('data/'+str(dataset) + '.txt')as f:
for j in f:
entry = [float(x) for x in j.split(" ")]
adj[int(entry[0]), int(entry[1])] = 1
adj[int(entry[1]), int(entry[0])] = 1
labels1 = np.loadtxt('data/'+str(dataset) + '_label.txt')
labels1 = labels1.astype(int)
labels1 = labels1[:,1:].flatten()
idx_train = np.loadtxt('data/'+str(dataset) + '_train_node.txt')
idx_train = idx_train.astype(int)
idx_val = np.loadtxt('data/'+str(dataset) + '_validation_node.txt')
idx_val = idx_val.astype(int)
idx_test = np.loadtxt('data/'+str(dataset) + '_test_node.txt')
idx_test = idx_test.astype(int)
features = np.eye(adj.shape[0])
elif dataset == 'cora_ml':
filename = 'data/' + str(dataset) + '_adj' + '.npz'
adj = sp.load_npz(filename)
filename = 'data/' + str(dataset) + '_features' + '.npz'
features = sp.load_npz(filename)
filename = 'data/' + str(dataset) + '_label' + '.npy'
labels1 = np.load(filename)
filename = 'data/' + str(dataset) + '_train_node' + '.npy'
idx_train = np.load(filename)
filename = 'data/' + str(dataset) + '_val_node' + '.npy'
idx_val = np.load(filename)
filename = 'data/' + str(dataset) + '_test_node' + '.npy'
idx_test = np.load(filename)
else:
filename = 'data/' + 'amazon_electronics_photo' + '_adj' + '.npz'
adj = sp.load_npz(filename)
filename = 'data/' + 'amazon_electronics_photo' + '_features' + '.npz'
features = sp.load_npz(filename)
filename = 'data/' + 'amazon_electronics_photo' + '_label' + '.npy'
labels1 = np.load(filename)
filename = 'data/' + 'amazon_electronics_photo'+ '_train_node' + '.npy'
idx_train = np.load(filename)
filename = 'data/' + 'amazon_electronics_photo' + '_val_node' + '.npy'
idx_val = np.load(filename)
filename = 'data/' + 'amazon_electronics_photo' + '_test_node' + '.npy'
idx_test = np.load(filename)
filename = 'data/' + 'amazon_electronics_photo' + '_label' + '.npy'
labels1 = np.load(filename)
return sp.csr_matrix(adj), sp.csr_matrix(features), idx_train, idx_val, idx_test, labels1
def get_adj( filename, require_lcc=True):
adj, features, labels = load_npz(filename)
adj = adj + adj.T
adj = adj.tolil()
adj[adj > 1] = 1
if require_lcc:
lcc = largest_connected_components(adj)
adj = adj[lcc][:, lcc]
features = features[lcc]
labels = labels[lcc]
assert adj.sum(0).A1.min() > 0, "Graph contains singleton nodes"
# whether to set diag=0?
adj.setdiag(0)
adj = adj.astype("float32").tocsr()
adj.eliminate_zeros()
assert np.abs(adj - adj.T).sum() == 0, "Input graph is not symmetric"
assert adj.max() == 1 and len(np.unique(adj[adj.nonzero()].A1)) == 1, "Graph must be unweighted"
return adj, features, labels
def load_npz(file_name, is_sparse=True):
with np.load(file_name) as loader:
# loader = dict(loader)
if is_sparse:
adj = sp.csr_matrix((loader['adj_data'], loader['adj_indices'],
loader['adj_indptr']), shape=loader['adj_shape'])
if 'attr_data' in loader:
features = sp.csr_matrix((loader['attr_data'], loader['attr_indices'],
loader['attr_indptr']), shape=loader['attr_shape'])
else:
features = None
labels = loader.get('labels')
else:
adj = loader['adj_data']
if 'attr_data' in loader:
features = loader['attr_data']
else:
features = None
labels = loader.get('labels')
if features is None:
features = np.eye(adj.shape[0])
features = sp.csr_matrix(features, dtype=np.float32)
return adj, features, labels
def get_train_val_test(nnodes, val_size=0.1, test_size=0.8, stratify=None, seed=None):
assert stratify is not None, 'stratify cannot be None!'
if seed is not None:
np.random.seed(seed)
idx = np.arange(nnodes)
train_size = 1 - val_size - test_size
idx_train_and_val, idx_test = train_test_split(idx,
random_state=None,
train_size=train_size + val_size,
test_size=test_size,
stratify=stratify)
if stratify is not None:
stratify = stratify[idx_train_and_val]
idx_train, idx_val = train_test_split(idx_train_and_val,
random_state=None,
train_size=(train_size / (train_size + val_size)),
test_size=(val_size / (train_size + val_size)),
stratify=stratify)
return idx_train, idx_val, idx_test
def largest_connected_components(adj, n_components=1):
_, component_indices = sp.csgraph.connected_components(adj)
component_sizes = np.bincount(component_indices)
components_to_keep = np.argsort(component_sizes)[::-1][:n_components] # reverse order to sort descending
nodes_to_keep = [
idx for (idx, component) in enumerate(component_indices) if component in components_to_keep]
print("Selecting {0} largest connected components".format(n_components))
return nodes_to_keep