def __init__(
self,
rw_hops=64,
subgraph_size=64,
restart_prob=0.8,
positional_embedding_size=32,
step_dist=[1.0, 0.0, 0.0],
):
super(GraphDataset).__init__()
self.rw_hops = rw_hops
self.subgraph_size = subgraph_size
self.restart_prob = restart_prob
self.positional_embedding_size = positional_embedding_size
self.step_dist = step_dist
assert sum(step_dist) == 1.0
assert positional_embedding_size > 1
# graphs = []
graphs, _ = dgl.data.utils.load_graphs(
"data_bin/dgl/lscc_graphs.bin", [0, 1, 2]
)
for name in ["cs", "physics"]:
g = Coauthor(name)[0]
g.remove_nodes((g.in_degrees() == 0).nonzero().squeeze())
g.readonly()
graphs.append(g)
for name in ["computers", "photo"]:
g = AmazonCoBuy(name)[0]
g.remove_nodes((g.in_degrees() == 0).nonzero().squeeze())
g.readonly()
graphs.append(g)
# more graphs are comming ...
print("load graph done")
self.graphs = graphs
self.length = sum([g.number_of_nodes() for g in self.graphs])
def extract_dataset():
parser = argparse.ArgumentParser(description='DATA')
register_data_args(parser)
args = parser.parse_args()
dataset_name = [
'cora', 'citeseer', 'pubmed', 'reddit', 'CoraFull', 'Coauthor_cs',
'Coauthor_physics', 'AmazonCoBuy_computers', 'AmazonCoBuy_photo'
]
print("Now PATH IS ", os.getcwd())
for name in dataset_name:
'''
if os.path.exists(name):
print('Folder exists. Skipping ' + name)
continue
'''
if name in ['cora', 'citeseer', 'pubmed', 'reddit']:
args.dataset = name
print('args.dataset = ', args.dataset)
if not os.path.exists(name):
os.mkdir(name)
os.chdir(name)
print("Now PATH IS ", os.getcwd())
data = load_data(args)
features = data.features
labels = data.labels
graph = data.graph
edges = graph.edges
train_mask = data.train_mask
val_mask = data.val_mask
test_mask = data.test_mask
n_nodes = features.shape[0]
n_edges = data.graph.number_of_edges
if args.dataset == 'reddit':
graph, features, labels, train_mask, val_mask, test_mask = cut_graph(
graph, n_nodes, n_edges, features, labels, train_mask,
val_mask, test_mask, 0.85)
#edge_x = np.append(edge_x, edge_y, axis=1)
edges_list = np.array([])
first_element = True
if name is not 'reddit':
for item in edges:
if first_element:
edges_list = np.array([[item[0], item[1]]])
first_element = False
else:
edges_list = np.append(edges_list,
np.array([[item[0], item[1]]]),
axis=0)
if name == 'reddit':
edges = graph.edges()
edge_x = edges[0].numpy().reshape((-1, 1))
print(edge_x.shape)
edge_y = edges[1].numpy().reshape((-1, 1))
edges_list = np.hstack((edge_x, edge_y))
print(edges_list.shape, edge_x.shape, edge_y.shape)
print('features_shape', features.shape)
print('labels_shape', labels.shape)
print('edges_shape', edges_list.shape)
'''
np.savetxt('edges.txt', edges_list)
np.savetxt('features.txt', features)
np.savetxt('labels.txt', labels)
np.savetxt('train_mask.txt', train_mask)
np.savetxt('val_mask.txt', val_mask)
np.savetxt('test_mask.txt', test_mask)
'''
np.save('edges.npy', edges_list)
np.save('features.npy', features)
np.save('labels.npy', labels)
np.save('train_mask.npy', train_mask)
print('Finish writing dataset', name)
os.chdir('..')
print('change to ', os.getcwd())
else:
if not os.path.exists(name):
os.mkdir(name)
os.chdir(name)
if name == 'CoraFull':
data = CoraFull()
elif name == 'Coauthor_cs':
data = Coauthor('cs')
elif name == 'Coauthor_physics':
data = Coauthor('physics')
elif name == 'AmazonCoBuy_computers':
data = AmazonCoBuy('computers')
elif name == 'AmazonCoBuy_photo':
data = AmazonCoBuy('photo')
else:
raise Exception("No such a dataset {}".format(name))
graph = data.data[0]
features = torch.FloatTensor(graph.ndata['feat']).numpy()
labels = torch.LongTensor(graph.ndata['label']).numpy()
print('dataset ', name)
features_shape = features.shape
labels_shape = labels.shape
n_nodes = features_shape[0]
edges_u, edges_v = graph.all_edges()
edges_u = edges_u.numpy()
edges_v = edges_v.numpy()
edges_list = np.array([])
first_element = True
for idx in range(len(edges_u)):
if first_element:
edges_list = np.array([[edges_u[idx], edges_v[idx]]])
first_element = False
else:
edges_list = np.append(edges_list,
np.array(
[[edges_u[idx], edges_v[idx]]]),
axis=0)
print('features_shape', features_shape)
print('labels_shape', labels_shape)
print('edges_shape', edges_list.shape)
train_mask = []
for x in range(500):
train_mask.append(True)
for x in range(n_nodes - 500):
train_mask.append(False)
train_mask = np.array(train_mask)
'''
np.savetxt('edges.txt', edges_list)
np.savetxt('features.txt', features)
np.savetxt('labels.txt', labels)
np.savetxt('train_mask.txt', train_mask)
'''
np.save('edges.npy', edges_list)
np.save('features.npy', features)
np.save('labels.npy', labels)
np.save('train_mask.npy', train_mask)
print('Finish writing dataset', name)
os.chdir('..')
print('change to ', os.getcwd())
def main(args):
torch.manual_seed(1234)
if args.dataset == 'cora' or args.dataset == 'citeseer' or args.dataset == 'pubmed':
data = load_data(args)
features = torch.FloatTensor(data.features)
labels = torch.LongTensor(data.labels)
in_feats = features.shape[1]
g = data.graph
if args.dataset == 'cora':
g.remove_edges_from(nx.selfloop_edges(g))
g.add_edges_from(zip(g.nodes(), g.nodes()))
g = DGLGraph(g)
attr_matrix = data.features
labels = data.labels
else:
if args.dataset == 'physics':
data = Coauthor('physics')
if args.dataset == 'cs':
data = Coauthor('cs')
if args.dataset == 'computers':
data = AmazonCoBuy('computers')
if args.dataset == 'photo':
data = AmazonCoBuy('photo')
g = data
g = data[0]
attr_matrix = g.ndata['feat']
labels = g.ndata['label']
features = torch.FloatTensor(g.ndata['feat'])
### LCC of the graph
n_components = 1
sparse_graph = g.adjacency_matrix_scipy(return_edge_ids=False)
_, component_indices = sp.csgraph.connected_components(sparse_graph)
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
]
adj_matrix = sparse_graph[nodes_to_keep][:, nodes_to_keep]
num_nodes = len(nodes_to_keep)
g = adj_matrix
g = DGLGraph(g)
g = remove_self_loop(g)
g = add_self_loop(g)
g = DGLGraph(g)
g.ndata['feat'] = attr_matrix[nodes_to_keep]
features = torch.FloatTensor(g.ndata['feat'].float())
if args.dataset == 'cora' or args.dataset == 'pubmed':
features = features / (features.norm(dim=1) + 1e-8)[:, None]
g.ndata['label'] = labels[nodes_to_keep]
labels = torch.LongTensor(g.ndata['label'])
in_feats = features.shape[1]
unique_l = np.unique(labels, return_counts=False)
n_classes = len(unique_l)
n_nodes = g.number_of_nodes()
n_edges = g.number_of_edges()
print('Number of nodes', n_nodes, 'Number of edges', n_edges)
enc = OneHotEncoder()
enc.fit(labels.reshape(-1, 1))
ylabels = enc.transform(labels.reshape(-1, 1)).toarray()
for beta in [args.beta]:
for K in [args.num_clusters]:
for alpha in [args.alpha]:
accs = []
t_st = time.time()
sets = "imbalanced"
for k in range(2): #number of differnet trainings
#print(k)
random_state = np.random.RandomState()
if sets == "imbalanced":
train_idx, val_idx, test_idx = get_train_val_test_split(
random_state,
ylabels,
train_examples_per_class=None,
val_examples_per_class=None,
test_examples_per_class=None,
train_size=20 * n_classes,
val_size=30 * n_classes,
test_size=None)
elif sets == "balanced":
train_idx, val_idx, test_idx = get_train_val_test_split(
random_state,
ylabels,
train_examples_per_class=20,
val_examples_per_class=30,
test_examples_per_class=None,
train_size=None,
val_size=None,
test_size=None)
else:
("No such set configuration (imbalanced/balanced)")
n_nodes = len(nodes_to_keep)
train_mask = np.zeros(n_nodes)
train_mask[train_idx] = 1
val_mask = np.zeros(n_nodes)
val_mask[val_idx] = 1
test_mask = np.zeros(n_nodes)
test_mask[test_idx] = 1
train_mask = torch.BoolTensor(train_mask)
val_mask = torch.BoolTensor(val_mask)
test_mask = torch.BoolTensor(test_mask)
"""
Planetoid Split for CORA, CiteSeer, PubMed
train_mask = torch.BoolTensor(data.train_mask)
val_mask = torch.BoolTensor(data.val_mask)
test_mask = torch.BoolTensor(data.test_mask)
train_mask2 = torch.BoolTensor(data.train_mask)
val_mask2 = torch.BoolTensor(data.val_mask)
test_mask2 = torch.BoolTensor(data.test_mask)
"""
if args.gpu < 0:
cuda = False
else:
cuda = True
torch.cuda.set_device(args.gpu)
features = features.cuda()
labels = labels.cuda()
train_mask = train_mask.cuda()
val_mask = val_mask.cuda()
test_mask = test_mask.cuda()
gic = GIC(g, in_feats, args.n_hidden, args.n_layers,
nn.PReLU(args.n_hidden), args.dropout, K, beta,
alpha)
if cuda:
gic.cuda()
gic_optimizer = torch.optim.Adam(
gic.parameters(),
lr=args.gic_lr,
weight_decay=args.weight_decay)
# train GIC
cnt_wait = 0
best = 1e9
best_t = 0
dur = []
for epoch in range(args.n_gic_epochs):
gic.train()
if epoch >= 3:
t0 = time.time()
gic_optimizer.zero_grad()
loss = gic(features)
#print(loss)
loss.backward()
gic_optimizer.step()
if loss < best:
best = loss
best_t = epoch
cnt_wait = 0
torch.save(gic.state_dict(), 'best_gic.pkl')
else:
cnt_wait += 1
if cnt_wait == args.patience:
#print('Early stopping!')
break
if epoch >= 3:
dur.append(time.time() - t0)
#print("Epoch {:05d} | Time(s) {:.4f} | Loss {:.4f} | "
#"ETputs(KTEPS) {:.2f}".format(epoch, np.mean(dur), loss.item(),
#n_edges / np.mean(dur) / 1000))
# train classifier
#print('Loading {}th epoch'.format(best_t))
gic.load_state_dict(torch.load('best_gic.pkl'))
embeds = gic.encoder(features, corrupt=False)
embeds = embeds / (embeds + 1e-8).norm(dim=1)[:, None]
embeds = embeds.detach()
# create classifier model
classifier = Classifier(args.n_hidden, n_classes)
if cuda:
classifier.cuda()
classifier_optimizer = torch.optim.Adam(
classifier.parameters(),
lr=args.classifier_lr,
weight_decay=args.weight_decay)
dur = []
best_a = 0
cnt_wait = 0
for epoch in range(args.n_classifier_epochs):
classifier.train()
if epoch >= 3:
t0 = time.time()
classifier_optimizer.zero_grad()
preds = classifier(embeds)
loss = F.nll_loss(preds[train_mask],
labels[train_mask])
loss.backward()
classifier_optimizer.step()
if epoch >= 3:
dur.append(time.time() - t0)
acc = evaluate(
classifier, embeds, labels, val_mask
) #+ evaluate(classifier, embeds, labels, train_mask)
if acc > best_a and epoch > 100:
best_a = acc
best_t = epoch
torch.save(classifier.state_dict(),
'best_class.pkl')
#print("Epoch {:05d} | Time(s) {:.4f} | Loss {:.4f} | Accuracy {:.4f} | "
#"ETputs(KTEPS) {:.2f}".format(epoch, np.mean(dur), loss.item(),
#acc, n_edges / np.mean(dur) / 1000))
acc = evaluate(classifier, embeds, labels, test_mask)
accs.append(acc)
print('=================== ', ' alpha', alpha, ' beta ', beta,
'K', K)
print(args.dataset, ' Acc (mean)', mean(accs), ' (std)',
stdev(accs))
print('=================== time', int(
(time.time() - t_st) / 60))