def test_from_matrix():
A = sp.csr_matrix(
np.array([[0, 2, 0, 3], [2, 0, 1, 1], [0, 1, 0, 0], [3, 1, 0, 0]]))
graph = preprocess.graph.Graph(A)
expected_edges = torch.tensor([(0, 1), (0, 3), (1, 2), (1, 3)])
assert graph.n_items == 4
assert graph.n_edges == 4
testing.assert_all_equal(expected_edges, graph.edges)
testing.assert_all_equal(np.array([2.0, 3.0, 1.0, 1.0]), graph.distances)
def test_graph_from_torch():
X = torch.ones(16).reshape(4, 4)
X.fill_diagonal_(0.0)
graph = preprocess.graph.Graph(X)
assert graph.n_items == 4
assert graph.n_edges == 6
testing.assert_all_equal(torch.ones(graph.n_edges), graph.distances)
expected_edges = torch.tensor([[0, 1], [0, 2], [0, 3], [1, 2], [1, 3],
[2, 3]])
testing.assert_all_equal(expected_edges, graph.edges)
def test_from_edges(device):
edges = torch.tensor([[0, 1], [1, 2], [2, 3], [3, 0]], device=device)
graph = preprocess.Graph.from_edges(edges)
assert graph.n_items == 4
assert graph.n_edges == edges.shape[0]
expected_edges = torch.tensor([[0, 1], [0, 3], [1, 2], [2, 3]])
testing.assert_all_equal(expected_edges, graph.edges)
expected_distances = torch.ones(graph.n_edges)
testing.assert_all_equal(expected_distances, graph.distances)
def test_graph_from_csc_matrix():
X = np.ones(16).reshape(4, 4)
np.fill_diagonal(X, 0.0)
X = sp.csc_matrix(X)
with testing.disable_logging():
# suppress a warning re: non-CSR matrix.
graph = preprocess.graph.Graph(X)
assert graph.n_items == 4
assert graph.n_edges == 6
testing.assert_all_equal(torch.ones(graph.n_edges), graph.distances)
expected_edges = torch.tensor([[0, 1], [0, 2], [0, 3], [1, 2], [1, 3],
[2, 3]])
testing.assert_all_equal(expected_edges, graph.edges)
def test_all_distances_numpy(device):
del device
np.random.seed(0)
data_matrix = np.random.randn(4, 2)
graph = preprocess.data_matrix.distances(data_matrix)
assert graph.n_items == data_matrix.shape[0]
assert graph.n_edges == 6
testing.assert_all_equal(
graph.edges,
torch.tensor([[0, 1], [0, 2], [0, 3], [1, 2], [1, 3], [2, 3]]),
)
for e, d in zip(graph.edges, graph.distances):
e = e.cpu().numpy()
d = d.item()
true_distance = np.linalg.norm(data_matrix[e[0]] - data_matrix[e[1]])
testing.assert_allclose(true_distance, d)
def test_all_distances_torch(device):
np.random.seed(0)
data_matrix = torch.tensor(np.random.randn(4, 2),
dtype=torch.float,
device=device)
graph = preprocess.data_matrix.distances(data_matrix)
assert graph.n_items == data_matrix.shape[0]
assert graph.n_edges == 6
testing.assert_all_equal(
graph.edges,
torch.tensor([[0, 1], [0, 2], [0, 3], [1, 2], [1, 3], [2, 3]]),
)
for e, d in zip(graph.edges, graph.distances):
e = e
d = d
true_distance = (data_matrix[e[0]] - data_matrix[e[1]]).norm()
testing.assert_allclose(true_distance, d)
def test_neighbors(device):
edges = torch.tensor([[0, 1], [1, 2], [2, 3], [3, 0]], device=device)
graph = preprocess.Graph.from_edges(edges)
neighbors = graph.neighbors(0)
testing.assert_all_equal(np.array([1, 3]), neighbors)
neighbors = graph.neighbors(1)
testing.assert_all_equal(np.array([0, 2]), neighbors)
neighbors = graph.neighbors(2)
testing.assert_all_equal(np.array([1, 3]), neighbors)
neighbors = graph.neighbors(3)
testing.assert_all_equal(np.array([0, 2]), neighbors)
def test_distances(device):
edges = torch.tensor([[0, 1], [1, 2], [2, 3], [3, 0]], device=device)
graph = preprocess.Graph.from_edges(edges,
torch.arange(edges.shape[0]) + 1)
distances = graph.neighbor_distances(0)
testing.assert_all_equal(np.array([1.0, 4.0]), distances)
distances = graph.neighbor_distances(1)
testing.assert_all_equal(np.array([1.0, 2.0]), distances)
distances = graph.neighbor_distances(2)
testing.assert_all_equal(np.array([2.0, 3.0]), distances)
distances = graph.neighbor_distances(3)
testing.assert_all_equal(np.array([4.0, 3.0]), distances)
def test_distances_repeated_edges_summed(device):
edges = torch.tensor([[0, 1], [1, 2], [2, 3], [3, 0], [0, 3]],
device=device)
graph = preprocess.Graph.from_edges(edges,
torch.arange(edges.shape[0]) + 1)
# edge (0, 3) repeated, so distance is 4 + 5 = 9
distances = graph.neighbor_distances(0)
testing.assert_all_equal(np.array([1.0, 9.0]), distances)
distances = graph.neighbor_distances(1)
testing.assert_all_equal(np.array([1.0, 2.0]), distances)
distances = graph.neighbor_distances(2)
testing.assert_all_equal(np.array([2.0, 3.0]), distances)
distances = graph.neighbor_distances(3)
testing.assert_all_equal(np.array([9.0, 3.0]), distances)
def test_adjacency_matrix(device):
edges = torch.tensor([[0, 1], [1, 2], [2, 3], [3, 0]], device=device)
graph = preprocess.Graph.from_edges(edges)
expected = np.array([[0, 1, 0, 1], [1, 0, 1, 0], [0, 1, 0, 1],
[1, 0, 1, 0]])
testing.assert_all_equal(expected, graph.A.todense())