def test_RGCN_apply_dense_directed():
G, features = create_graph_features(is_directed=True)
As = get_As(G)
As = [np.expand_dims(A.todense(), 0) for A in As]
generator = RelationalFullBatchNodeGenerator(G, sparse=False)
rgcnModel = RGCN([2],
generator,
num_bases=10,
activations=["relu"],
dropout=0.5)
x_in, x_out = rgcnModel.in_out_tensors()
model = keras.Model(inputs=x_in, outputs=x_out)
# Check fit method
out_indices = np.array([[0, 1]], dtype="int32")
preds_1 = model.predict([features[None, :, :], out_indices] + As)
assert preds_1.shape == (1, 2, 2)
# Check fit method
preds_2 = model.predict(generator.flow(["a", "b"]))
assert preds_2.shape == (1, 2, 2)
assert preds_1 == pytest.approx(preds_2)
def generator_flow(self, G, node_ids, node_targets, sparse=False):
generator = RelationalFullBatchNodeGenerator(G, sparse=sparse)
n_nodes = len(G.nodes())
gen = generator.flow(node_ids, node_targets)
if sparse:
[X, tind, *As], y = gen[0]
As_indices = As[:self.num_relationships]
As_values = As[self.num_relationships:]
As_sparse = [
sps.coo_matrix(
(A_val[0], (A_ind[0, :, 0], A_ind[0, :, 1])),
shape=(n_nodes, n_nodes),
) for A_ind, A_val in zip(As_indices, As_values)
]
As_dense = [A.toarray() for A in As_sparse]
else:
[X, tind, *As], y = gen[0]
As_dense = As
assert np.allclose(X,
gen.features) # X should be equal to gen.features
assert tind.shape[1] == len(node_ids)
if node_targets is not None:
assert np.allclose(y, node_targets)
return As_dense, tind, y
def test_RGCN_apply_sparse_directed():
G, features = create_graph_features(is_directed=True)
As = get_As(G)
As = [A.tocoo() for A in As]
A_indices = [
np.expand_dims(np.hstack((A.row[:, None], A.col[:, None])), 0) for A in As
]
A_values = [np.expand_dims(A.data, 0) for A in As]
generator = RelationalFullBatchNodeGenerator(G, sparse=True)
rgcnModel = RGCN([2], generator, num_bases=10, activations=["relu"], dropout=0.5)
x_in, x_out = rgcnModel.build()
model = keras.Model(inputs=x_in, outputs=x_out)
# Check fit method
out_indices = np.array([[0, 1]], dtype="int32")
preds_1 = model.predict([features[None, :, :], out_indices] + A_indices + A_values)
assert preds_1.shape == (1, 2, 2)
# Check fit_generator method
preds_2 = model.predict_generator(generator.flow(["a", "b"]))
assert preds_2.shape == (1, 2, 2)
assert preds_1 == pytest.approx(preds_2)
def test_generator_flow_targets_as_list(self):
generator = RelationalFullBatchNodeGenerator(self.G)
node_ids = list(self.G.nodes())[:3]
node_targets = [1] * len(node_ids)
gen = generator.flow(node_ids, node_targets)
inputs, y = gen[0]
assert y.shape == (1, 3)
assert np.sum(y) == 3
def test_generator_constructor(self):
generator = RelationalFullBatchNodeGenerator(self.G)
assert len(generator.As) == self.num_relationships
assert all((A.shape == (self.N, self.N)) for A in generator.As)
assert generator.features.shape == (self.N, self.n_feat)
def test_fullbatch_generator_transform(self):
G, _ = create_graph_features()
def func(features, A, **kwargs):
return features, A.dot(A)
generator = RelationalFullBatchNodeGenerator(G, "test", transform=func)
assert generator.name == "test"
As = []
edge_types = sorted(set(e[-1]
for e in G.edges(include_edge_type=True)))
node_list = list(G.nodes())
node_index = dict(zip(node_list, range(len(node_list))))
for edge_type in edge_types:
col_index = [
node_index[n1]
for n1, n2, etype in G.edges(include_edge_type=True)
if etype == edge_type
]
row_index = [
node_index[n2]
for n1, n2, etype in G.edges(include_edge_type=True)
if etype == edge_type
]
data = np.ones(len(col_index), np.float64)
A = sps.coo_matrix((data, (row_index, col_index)),
shape=(len(node_list), len(node_list)))
As.append(A)
assert all(
np.array_equal(A_1.dot(A_1).todense(), A_2.todense())
for A_1, A_2 in zip(As, generator.As))
def test_fullbatch_generator_init_3(self):
G, _ = create_graph_features()
func = "Not callable"
with pytest.raises(TypeError):
generator = RelationalFullBatchNodeGenerator(G, "test", transform=func)
def test_RGCN_init():
G, features = create_graph_features()
generator = RelationalFullBatchNodeGenerator(G)
rgcnModel = RGCN([2], generator, num_bases=10, activations=["relu"], dropout=0.5)
assert rgcnModel.layer_sizes == [2]
assert rgcnModel.activations == ["relu"]
assert rgcnModel.dropout == 0.5
assert rgcnModel.num_bases == 10
def test_weighted(self):
G, _ = create_graph_features(edge_weights=True)
generator = RelationalFullBatchNodeGenerator(G,
weighted=True,
transform=lambda f, A:
(f, A))
np.testing.assert_array_equal(
generator.As[0].todense(),
[[0, 2.0, 0], [2.0, 0, 0.5], [0, 0.5, 0]])
np.testing.assert_array_equal(
generator.As[1].todense(),
[[0, 0.0, 1.0], [0.0, 0, 0.0], [1.0, 0.0, 0]])
def test_kernel_and_bias_defaults():
graph, _ = create_graph_features()
generator = RelationalFullBatchNodeGenerator(graph, sparse=False)
rgcn = RGCN([2, 2], generator, num_bases=10)
for layer in rgcn._layers:
if isinstance(layer, RelationalGraphConvolution):
assert isinstance(layer.kernel_initializer,
tf.initializers.GlorotUniform)
assert isinstance(layer.bias_initializer, tf.initializers.Zeros)
assert layer.kernel_regularizer is None
assert layer.bias_regularizer is None
assert layer.kernel_constraint is None
assert layer.bias_constraint is None
def test_generator_constructor_wrong_G_type(self):
with pytest.raises(TypeError):
generator = RelationalFullBatchNodeGenerator(self.G.to_networkx())
def test_fullbatch_generator_init_1(self):
G, feats = create_graph_features()
generator = RelationalFullBatchNodeGenerator(G, name="test")
assert generator.name == "test"
assert np.array_equal(feats, generator.features)
def test_generator_flow_targets_not_iterator(self):
generator = RelationalFullBatchNodeGenerator(self.G)
node_ids = list(self.G.nodes())[:3]
node_targets = 1
with pytest.raises(TypeError):
generator.flow(node_ids, node_targets)
def test_fullbatch_generator_init_heterogeneous_nodes(self):
G = example_hin_1(feature_sizes={})
with pytest.raises(ValueError,
match="G: expected one node type, found 'A', 'B'"):
RelationalFullBatchNodeGenerator(G)
def test_RGCN_save_load(tmpdir, num_bases, sparse):
graph, _ = create_graph_features()
generator = RelationalFullBatchNodeGenerator(graph, sparse=sparse)
rgcn = RGCN([2, 2], generator, num_bases=num_bases)
test_utils.model_save_load(tmpdir, rgcn)
