def test_points_mapping():
points = np.array([[0, 0, 0], [10, 10, 10], [1000, 1000, 1000]])
source_space = AnatomicalSpace(
"psl", shape=(1000, 1000, 1000), resolution=(1, 1, 1)
)
target_space = AnatomicalSpace(
"asl", shape=(100, 100, 100), resolution=(10, 10, 10)
)
mapped_points = source_space.map_points_to(target_space, points)
assert np.allclose(
mapped_points,
np.array([[100.0, 0.0, 0.0], [99.0, 1.0, 1.0], [0.0, 100.0, 100.0]]),
)
def test_point_transform(src_o, tgt_o, src_shape, tgt_shape):
# Create an array with descriptive space labels:
source_stack = create_label_array(src_o, src_shape)
# Create spaces objects:
source_space = AnatomicalSpace(src_o, shape=[s * 2 for s in src_shape])
# Test both mapping to AnatomicalSpace obj and origin with decorator:
for target_space in [tgt_o, AnatomicalSpace(tgt_o)]:
# Define grid of points sampling 4 points per axis:
grid_positions = [[1, s - 1, s + 1, s * 2 - 1] for s in src_shape]
source_pts = np.array(list(itertools.product(*grid_positions)))
# Map points and stack to target space:
mapped_pts = source_space.map_points_to(target_space, source_pts)
mapped_stack = source_space.map_stack_to(target_space, source_stack)
# Check that point coordinates keep the same values:
for p_source, p_mapped in zip(source_pts, mapped_pts):
p_s, p_m = [tuple(p.astype(np.int)) for p in [p_source, p_mapped]]
assert source_stack[p_s] == mapped_stack[p_m]
def test_point_transform_fail():
s = AnatomicalSpace("asl")
with pytest.raises(TypeError) as error:
s.map_points_to("psl", np.array([[0, 1, 2], [0, 1, 2]]))
assert "The source space should have a shape" in str(error)