def test__3x3_grid_of_pixel_grid__1_coordinate_per_square_pixel__near_edges_of_pixels(
self):
# _ _ _
# |_|_|_| Boundaries for pixels x = 0 and y = 0 -1.0 to -(1/3)
# |_|_|_| Boundaries for pixels x = 1 and y = 1 - (1/3) to (1/3)
# |_|_|_| Boundaries for pixels x = 2 and y = 2 - (1/3)" to 1.0"
pixelization_grid = np.array([[1.0, -1.0], [1.0, 0.0], [1.0, 1.0],
[-0.32, -1.0], [-0.32, 0.32],
[0.0, 1.0], [-0.34, -0.34],
[-0.34, 0.325], [-1.0, 1.0]])
pix = pixelizations.Rectangular(shape=(3, 3))
geometry = pix.geometry_from_grid(grid=pixelization_grid)
grids = MockGridStack(regular=pixelization_grid,
sub=MockSubGrid(pixelization_grid,
sub_to_regular=np.ones((1)),
sub_grid_size=1))
pix = mappers.RectangularMapper(pixels=9,
shape=(3, 3),
grid_stack=grids,
border=None,
geometry=geometry)
assert (pix.regular_to_pix == np.array([0, 1, 2, 3, 4, 5, 6, 7,
8])).all()
assert (pix.sub_to_pix == np.array([0, 1, 2, 3, 4, 5, 6, 7,
8])).all()
assert pix.pix_to_regular == [[0], [1], [2], [3], [4], [5], [6],
[7], [8]]
assert pix.pix_to_sub == [[0], [1], [2], [3], [4], [5], [6], [7],
[8]]
def test__4x3_grid__non_symmetric_centre_shift(self):
# _ _ _
# |_|_|_|
# |_|_|_|
# |_|_|_|
# |_|_|_|
pixelization_grid = np.array([[3.0, -0.5], [3.0,
0.51], [2.49, -0.5],
[1.4, 1.0], [1.0, 2.5]])
pix = pixelizations.Rectangular(shape=(4, 3))
geometry = pix.geometry_from_grid(grid=pixelization_grid)
grids = MockGridStack(regular=pixelization_grid,
sub=MockSubGrid(pixelization_grid,
sub_to_regular=np.ones((1)),
sub_grid_size=1))
pix = mappers.RectangularMapper(pixels=12,
shape=(4, 3),
grid_stack=grids,
border=None,
geometry=geometry)
assert (pix.regular_to_pix == np.array([0, 1, 3, 10, 11])).all()
assert (pix.sub_to_pix == np.array([0, 1, 3, 10, 11])).all()
assert pix.pix_to_regular == [[0], [1], [], [2], [], [], [], [],
[], [], [3], [4]]
assert pix.pix_to_sub == [[0], [1], [], [2], [], [], [], [], [],
[], [3], [4]]
def test__3x3_grid__buffer_is_small__grid_give_min_minus_1_max_1__sets_up_geometry_correctly(
self):
pix = pixelizations.Rectangular(shape=(3, 3))
pixelization_grid = np.array([[1.0, -1.0], [1.0, 0.0], [1.0, 1.0],
[0.0, -1.0], [0.0, 0.0], [0.0, 1.0],
[-1.0, -1.0], [-1.0, 0.0],
[-1.0, 1.0]])
geometry = pix.geometry_from_grid(pixelization_grid, buffer=1e-8)
assert geometry.shape == (3, 3)
assert geometry.pixel_scales == pytest.approx((2. / 3., 2. / 3.),
1e-2)
assert (geometry.pixel_neighbors[0] == [1, 3, -1, -1]).all()
assert (geometry.pixel_neighbors[1] == [0, 2, 4, -1]).all()
assert (geometry.pixel_neighbors[2] == [1, 5, -1, -1]).all()
assert (geometry.pixel_neighbors[3] == [0, 4, 6, -1]).all()
assert (geometry.pixel_neighbors[4] == [1, 3, 5, 7]).all()
assert (geometry.pixel_neighbors[5] == [2, 4, 8, -1]).all()
assert (geometry.pixel_neighbors[6] == [3, 7, -1, -1]).all()
assert (geometry.pixel_neighbors[7] == [4, 6, 8, -1]).all()
assert (geometry.pixel_neighbors[8] == [5, 7, -1, -1]).all()
assert (geometry.pixel_neighbors_size == np.array(
[2, 3, 2, 3, 4, 3, 2, 3, 2])).all()
def test__3x3_grid__change_arcsecond_dimensions__not_symmetric(self):
# _ _ _
# |_|_|_| Boundaries for pixels x = 0 and y = 0 -1.5 to -0.5
# |_|_|_| Boundaries for pixels x = 1 and y = 1 -0.5 to 0.5
# |_|_|_| Boundaries for pixels x = 2 and y = 2 0.5 to 1.5
pixelization_grid = np.array([[1.0, -1.5], [1.0, -0.49],
[0.32, -1.5], [0.32, 0.51],
[-1.0, 1.5]])
pix = pixelizations.Rectangular(shape=(3, 3))
geometry = pix.geometry_from_grid(grid=pixelization_grid)
grids = MockGridStack(regular=pixelization_grid,
sub=MockSubGrid(pixelization_grid,
sub_to_regular=np.ones((1)),
sub_grid_size=1))
pix = mappers.RectangularMapper(pixels=9,
shape=(3, 3),
grid_stack=grids,
border=None,
geometry=geometry)
assert (pix.regular_to_pix == np.array([0, 1, 3, 5, 8])).all()
assert (pix.sub_to_pix == np.array([0, 1, 3, 5, 8])).all()
assert pix.pix_to_regular == [[0], [1], [], [2], [], [3], [], [],
[4]]
assert pix.pix_to_sub == [[0], [1], [], [2], [], [3], [], [], [4]]
def test__3x3_grid_of_pixel_grid__add_multiple_grid_to_1_pixel_pixel(
self):
# _ _ _
# -1.0 to -(1/3) |_|_|_|
# -(1/3) to (1/3) |_|_|_|
# (1/3) to 1.0 |_|_|_|
pixelization_grid = np.array([[1.0, -1.0], [0.0, 0.0], [1.0, 1.0],
[0.0, 0.0], [0.0, 0.0], [0.0, 0.0],
[-1.0, -1.0], [0.0, 0.0],
[-1.0, 1.0]])
pix = pixelizations.Rectangular(shape=(3, 3))
geometry = pix.geometry_from_grid(grid=pixelization_grid)
grids = MockGridStack(regular=pixelization_grid,
sub=MockSubGrid(pixelization_grid,
sub_to_regular=np.ones((1)),
sub_grid_size=1))
pix = mappers.RectangularMapper(pixels=9,
shape=(3, 3),
grid_stack=grids,
border=None,
geometry=geometry)
assert (pix.regular_to_pix == np.array([0, 4, 2, 4, 4, 4, 6, 4,
8])).all()
assert (pix.sub_to_pix == np.array([0, 4, 2, 4, 4, 4, 6, 4,
8])).all()
assert pix.pix_to_regular == [[0], [], [2], [], [1, 3, 4, 5, 7],
[], [6], [], [8]]
assert pix.pix_to_sub == [[0], [], [2], [], [1, 3, 4, 5, 7], [],
[6], [], [8]]
def test__4x3_grid__pixel_centres(self):
pix = pixelizations.Rectangular(shape=(4, 3))
pixelization_grid = np.array([[1.0, -1.0], [1.0, 0.0], [1.0, 1.0],
[0.0, -1.0], [0.0, 0.0], [0.0, 1.0],
[-1.0, -1.0], [-1.0, 0.0],
[-1.0, 1.0]])
geometry = pix.geometry_from_grid(pixelization_grid, buffer=1e-8)
assert geometry.pixel_centres == pytest.approx(
np.array([
[0.75, -2. / 3.],
[0.75, 0.0],
[0.75, 2. / 3.],
[0.25, -2. / 3.],
[0.25, 0.0],
[0.25, 2. / 3.],
[-0.25, -2. / 3.],
[-0.25, 0.0],
[-0.25, 2. / 3.],
[-0.75, -2. / 3.],
[-0.75, 0.0],
[-0.75, 2. / 3.],
]))
def test__3x3_pixelization__solution_vector_ascending(self):
pixelization_grid = np.array([[2.0, -1.0], [2.0, 0.0], [2.0, 1.0],
[0.0, -1.0], [0.0, 0.0], [0.0, 1.0],
[-2.0, -1.0], [-2.0, 0.0],
[-2.0, 1.0]])
pix = pixelizations.Rectangular(shape=(3, 3))
geometry = pix.geometry_from_grid(grid=pixelization_grid)
pix = mappers.RectangularMapper(pixels=9,
shape=(3, 3),
grid_stack=None,
border=None,
geometry=geometry)
recon_pix = pix.reconstructed_pixelization_from_solution_vector(
solution_vector=np.array(
[1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0]))
assert (recon_pix == np.array([[1.0, 2.0, 3.0], [4.0, 5.0, 6.0],
[7.0, 8.0, 9.0]])).all()
assert recon_pix.pixel_scales == pytest.approx((4. / 3., 2. / 3.),
1e-2)
assert recon_pix.origin == (0.0, 0.0)
def test__3x4_grid__change_arcsecond_dimensions__not_symmetric(self):
# _ _ _ _
# |_|_|_|_|
# |_|_|_|_|
# |_|_|_|_|
pixelization_grid = np.array([[1.0, -1.5], [1.0, -0.49],
[0.32, -1.5], [-0.34, 0.49],
[-1.0, 1.5]])
pix = pixelizations.Rectangular(shape=(3, 4))
grids = MockGridStack(regular=pixelization_grid,
sub=MockSubGrid(pixelization_grid,
sub_to_regular=np.ones((1)),
sub_grid_size=1))
geometry = pix.geometry_from_grid(grid=pixelization_grid)
pix = mappers.RectangularMapper(pixels=12,
shape=(3, 4),
grid_stack=grids,
border=None,
geometry=geometry)
assert (pix.regular_to_pix == np.array([0, 1, 4, 10, 11])).all()
assert (pix.sub_to_pix == np.array([0, 1, 4, 10, 11])).all()
assert pix.pix_to_regular == [[0], [1], [], [], [2], [], [], [],
[], [], [3], [4]]
assert pix.pix_to_sub == [[0], [1], [], [], [2], [], [], [], [],
[], [3], [4]]
def test__5_simple_grid__include_sub_grid(self):
# Source-plane comprises 5 grid, so 5 masked_image pixels traced to the pix-plane.
regular_grid = np.array([[1.0, -1.0], [1.0, 1.0], [0.0, 0.0],
[-1.0, -1.0], [-1.0, 1.0]])
# Assume a 2x2 sub-grid, so each of our 5 masked_image-pixels are split into 4.
# The grid below is unphysical in that the (0.0, 0.0) terms on the end of each sub-grid probably couldn't
# happen for a real lensing calculation. This is to make a mapping_matrix matrix which explicitly tests the
# sub-grid.
sub_grid = np.array([[1.0, -1.0], [1.0, -1.0], [1.0, -1.0], [0.0, 0.0],
[1.0, 1.0], [1.0, 1.0], [1.0, 1.0], [0.0, 0.0],
[0.0, 0.0], [0.0, 0.0], [0.0, 0.0], [0.0, 0.0],
[-1.0, -1.0], [-1.0, -1.0], [-1.0, -1.0],
[0.0, 0.0], [-1.0, 1.0], [-1.0, 1.0], [-1.0, 1.0],
[0.0, 0.0]])
sub_to_regular = np.array(
[0, 0, 0, 2, 1, 1, 1, 2, 2, 2, 2, 2, 3, 3, 3, 2, 4, 4, 4, 2])
grid_stack = MockGridStack(regular=regular_grid,
sub=MockSubGrid(sub_grid,
sub_to_regular,
sub_grid_size=2))
pix = pixelizations.Rectangular(shape=(3, 3))
mapper = pix.mapper_from_grid_stack_and_border(grid_stack=grid_stack,
border=None)
assert mapper.is_image_plane_pixelization == False
assert mapper.geometry.shape_arc_seconds == pytest.approx((2.0, 2.0),
1.0e-4)
assert mapper.geometry.origin == pytest.approx((0.0, 0.0), 1.0e-4)
assert (mapper.mapping_matrix == np.array(
[[0.75, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0],
[0.0, 0.0, 0.75, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0],
[0.0, 0.0, 0.0, 0.0, 2.0, 0.0, 0.0, 0.0, 0.0],
[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.75, 0.0, 0.0],
[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.75]])).all()
assert mapper.shape == (3, 3)
reg = regularization.Constant(coefficients=(1.0, ))
regularization_matrix = reg.regularization_matrix_from_pixel_neighbors(
mapper.geometry.pixel_neighbors,
mapper.geometry.pixel_neighbors_size)
assert (regularization_matrix == np.array(
[[2.00000001, -1.0, 0.0, -1.0, 0.0, 0.0, 0.0, 0.0, 0.0],
[-1.0, 3.00000001, -1.0, 0.0, -1.0, 0.0, 0.0, 0.0, 0.0],
[0.0, -1.0, 2.00000001, 0.0, 0.0, -1.0, 0.0, 0.0, 0.0],
[-1.0, 0.0, 0.0, 3.00000001, -1.0, 0.0, -1.0, 0.0, 0.0],
[0.0, -1.0, 0.0, -1.0, 4.00000001, -1.0, 0.0, -1.0, 0.0],
[0.0, 0.0, -1.0, 0.0, -1.0, 3.00000001, 0.0, 0.0, -1.0],
[0.0, 0.0, 0.0, -1.0, 0.0, 0.0, 2.00000001, -1.0, 0.0],
[0.0, 0.0, 0.0, 0.0, -1.0, 0.0, -1.0, 3.00000001, -1.0],
[0.0, 0.0, 0.0, 0.0, 0.0, -1.0, 0.0, -1.0, 2.00000001]])).all()
def test_fixed_pixelization(self):
galaxy_prior = gp.GalaxyModel(variable_redshift=True,
pixelization=pixelizations.Rectangular(),
regularization=regularization.Constant())
arguments = {galaxy_prior.redshift.redshift: 2.0}
galaxy = galaxy_prior.instance_for_arguments(arguments)
assert galaxy.pixelization.shape[0] == 3
assert galaxy.pixelization.shape[1] == 3
def test__3x3_grid__larger_range_of_grid(self):
pix = pixelizations.Rectangular(shape=(3, 3))
pixelization_grid = np.array([[2.0, 1.0], [4.0, 3.0], [6.0, 5.0],
[8.0, 7.0]])
geometry = pix.geometry_from_grid(pixelization_grid, buffer=1e-8)
assert geometry.shape == (3, 3)
assert geometry.pixel_scales == pytest.approx((6. / 3., 6. / 3.),
1e-2)
def test__3x3_grid__same_as_above_change_buffer(self):
pix = pixelizations.Rectangular(shape=(3, 3))
pixelization_grid = np.array([[-1.0, -1.0], [-1.0, 0.0],
[-1.0, 1.0], [0.0, -1.0], [0.0, 0.0],
[0.0, 1.0], [1.0, -1.0], [1.0, 0.0],
[1.0, 1.0]])
geometry = pix.geometry_from_grid(pixelization_grid, buffer=1e-4)
assert geometry.shape == (3, 3)
assert geometry.pixel_scales == pytest.approx((2. / 3., 2. / 3.),
1e-2)
def test__5_simple_grid__no_sub_grid(self):
# Source-plane comprises 5 grid, so 5 masked_image pixels traced to the pix-plane.
regular_grid = np.array([[1.0, -1.0], [1.0, 1.0], [0.0, 0.0],
[-1.0, -1.0], [-1.0, 1.0]])
sub_grid = np.array([[1.0, -1.0], [1.0, 1.0], [0.0, 0.0], [-1.0, -1.0],
[-1.0, 1.0]])
sub_to_regular = np.array([0, 1, 2, 3, 4])
grid_stack = MockGridStack(regular=regular_grid,
sub=MockSubGrid(
sub_grid=sub_grid,
sub_to_regular=sub_to_regular,
sub_grid_size=1))
# There is no sub-grid, so our sub_grid are just the masked_image grid (note the NumPy weighted_data structure
# ensures this has no sub-gridding)
pix = pixelizations.Rectangular(shape=(3, 3))
mapper = pix.mapper_from_grid_stack_and_border(grid_stack=grid_stack,
border=None)
assert mapper.is_image_plane_pixelization == False
assert mapper.geometry.shape_arc_seconds == pytest.approx((2.0, 2.0),
1.0e-4)
assert mapper.geometry.origin == pytest.approx((0.0, 0.0), 1.0e-4)
assert (mapper.mapping_matrix == np.array(
[[1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0],
[0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0],
[0.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 0.0],
[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0],
[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0]])).all()
assert mapper.shape == (3, 3)
reg = regularization.Constant(coefficients=(1.0, ))
regularization_matrix = reg.regularization_matrix_from_pixel_neighbors(
mapper.geometry.pixel_neighbors,
mapper.geometry.pixel_neighbors_size)
assert (regularization_matrix == np.array(
[[2.00000001, -1.0, 0.0, -1.0, 0.0, 0.0, 0.0, 0.0, 0.0],
[-1.0, 3.00000001, -1.0, 0.0, -1.0, 0.0, 0.0, 0.0, 0.0],
[0.0, -1.0, 2.00000001, 0.0, 0.0, -1.0, 0.0, 0.0, 0.0],
[-1.0, 0.0, 0.0, 3.00000001, -1.0, 0.0, -1.0, 0.0, 0.0],
[0.0, -1.0, 0.0, -1.0, 4.00000001, -1.0, 0.0, -1.0, 0.0],
[0.0, 0.0, -1.0, 0.0, -1.0, 3.00000001, 0.0, 0.0, -1.0],
[0.0, 0.0, 0.0, -1.0, 0.0, 0.0, 2.00000001, -1.0, 0.0],
[0.0, 0.0, 0.0, 0.0, -1.0, 0.0, -1.0, 3.00000001, -1.0],
[0.0, 0.0, 0.0, 0.0, 0.0, -1.0, 0.0, -1.0, 2.00000001]])).all()
def test__5x4_grid__buffer_is_small(self):
pix = pixelizations.Rectangular(shape=(5, 4))
pixelization_grid = np.array([[1.0, -1.0], [1.0, 0.0], [1.0, 1.0],
[0.0, -1.0], [0.0, 0.0], [0.0, 1.0],
[-1.0, -1.0], [-1.0, 0.0],
[-1.0, 1.0]])
geometry = pix.geometry_from_grid(pixelization_grid, buffer=1e-8)
assert geometry.shape == (5, 4)
assert geometry.pixel_scales == pytest.approx((2. / 5., 2. / 4.),
1e-2)
def test__compare_to_pixelization_util(self):
# |0 | 1| 2| 3|
# |4 | 5| 6| 7|
# |8 | 9|10|11|
# |12|13|14|15|
pix = pixelizations.Rectangular(shape=(7, 5))
pixel_neighbors, pixel_neighbors_size = pix.neighbors_from_pixelization(
)
pixel_neighbors_util, pixel_neighbors_size_util = \
pixelization_util.rectangular_neighbors_from_shape(shape=(7, 5))
assert (pixel_neighbors == pixel_neighbors_util).all()
assert (pixel_neighbors_size == pixel_neighbors_size_util).all()
def test__grid__requires_border_relocation(self):
regular_grid = np.array([[1.0, -1.0], [1.0, 1.0], [0.0, 0.0],
[-1.0, -1.0], [-1.0, 1.0]])
sub_grid = np.array([[2.0, 2.0], [2.0, 2.0], [2.0, 2.0], [2.0, 2.0],
[-2.0, -2.0]])
border = grids.RegularGridBorder(arr=np.array([0, 1, 3, 4]))
sub_to_regular = np.array([0, 1, 2, 3, 4])
grid_stack = MockGridStack(regular=regular_grid,
sub=MockSubGrid(sub_grid,
sub_to_regular,
sub_grid_size=1))
pix = pixelizations.Rectangular(shape=(3, 3))
mapper = pix.mapper_from_grid_stack_and_border(grid_stack, border)
assert mapper.is_image_plane_pixelization == False
assert mapper.geometry.shape_arc_seconds == pytest.approx((2.0, 2.0),
1.0e-4)
assert mapper.geometry.origin == pytest.approx((0.0, 0.0), 1.0e-4)
assert (mapper.mapping_matrix == np.array(
[[0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0],
[0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0],
[0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0],
[0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0],
[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0]])).all()
assert mapper.shape == (3, 3)
reg = regularization.Constant(coefficients=(1.0, ))
regularization_matrix = reg.regularization_matrix_from_pixel_neighbors(
mapper.geometry.pixel_neighbors,
mapper.geometry.pixel_neighbors_size)
assert (regularization_matrix == np.array(
[[2.00000001, -1.0, 0.0, -1.0, 0.0, 0.0, 0.0, 0.0, 0.0],
[-1.0, 3.00000001, -1.0, 0.0, -1.0, 0.0, 0.0, 0.0, 0.0],
[0.0, -1.0, 2.00000001, 0.0, 0.0, -1.0, 0.0, 0.0, 0.0],
[-1.0, 0.0, 0.0, 3.00000001, -1.0, 0.0, -1.0, 0.0, 0.0],
[0.0, -1.0, 0.0, -1.0, 4.00000001, -1.0, 0.0, -1.0, 0.0],
[0.0, 0.0, -1.0, 0.0, -1.0, 3.00000001, 0.0, 0.0, -1.0],
[0.0, 0.0, 0.0, -1.0, 0.0, 0.0, 2.00000001, -1.0, 0.0],
[0.0, 0.0, 0.0, 0.0, -1.0, 0.0, -1.0, 3.00000001, -1.0],
[0.0, 0.0, 0.0, 0.0, 0.0, -1.0, 0.0, -1.0, 2.00000001]])).all()
def test__setup_pixelization__galaxies_have_other_pixelization__returns_normal_grids(
self):
ma = mask.Mask(np.array([[False, False, False], [False, False, False],
[False, False, False]]),
pixel_scale=1.0)
grid_stack = grids.GridStack.grid_stack_from_mask_sub_grid_size_and_psf_shape(
mask=ma, sub_grid_size=1, psf_shape=(1, 1))
galaxy = g.Galaxy(pixelization=pixelizations.Rectangular(shape=(3, 3)),
regularization=regularization.Constant())
image_plane_pix_grids = \
pixelizations.setup_image_plane_pixelization_grid_from_galaxies_and_grid_stack(galaxies=[galaxy, galaxy],
grid_stack=grid_stack)
assert image_plane_pix_grids == grid_stack
def test__4x3_grid_of_pixel_grid__1_coordinate_in_each_pixel(self):
# _ _ _
# |_|_|_|
# |_|_|_|
# |_|_|_|
# |_|_|_|
# Boundaries for column pixel 0 -1.0 to -(1/3)
# Boundaries for column pixel 1 -(1/3) to (1/3)
# Boundaries for column pixel 2 (1/3) to 1.0
# Bounadries for row pixel 0 -1.0 to -0.5
# Bounadries for row pixel 1 -0.5 to 0.0
# Bounadries for row pixel 2 0.0 to 0.5
# Bounadries for row pixel 3 0.5 to 1.0
pixelization_grid = np.array([[1.0, -1.0], [1.0, 0.0], [1.0, 1.0],
[0.5, -1.0], [-0.5, 1.0],
[-1.0, 1.0]])
pix = pixelizations.Rectangular(shape=(4, 3))
geometry = pix.geometry_from_grid(grid=pixelization_grid)
grids = MockGridStack(regular=pixelization_grid,
sub=MockSubGrid(pixelization_grid,
sub_to_regular=np.ones((1)),
sub_grid_size=1))
pix = mappers.RectangularMapper(pixels=12,
shape=(4, 3),
grid_stack=grids,
border=None,
geometry=geometry)
assert (pix.regular_to_pix == np.array([0, 1, 2, 3, 8, 11])).all()
assert (pix.sub_to_pix == np.array([0, 1, 2, 3, 8, 11])).all()
assert pix.pix_to_regular == [[0], [1], [2], [3], [], [], [], [],
[4], [], [], [5]]
assert pix.pix_to_sub == [[0], [1], [2], [3], [], [], [], [], [4],
[], [], [5]]
def test__fit_figure_of_merit__matches_correct_fit_given_galaxy_profiles(
self, ccd_data):
lens_galaxy = g.Galaxy(light=lp.EllipticalSersic(intensity=0.1))
source_galaxy = g.Galaxy(
pixelization=pix.Rectangular(shape=(4, 4)),
regularization=reg.Constant(coefficients=(1.0, )))
phase = ph.LensPlanePhase(lens_galaxies=[lens_galaxy],
mask_function=ph.default_mask_function,
cosmology=cosmo.FLRW,
phase_name='test_phase')
analysis = phase.make_analysis(data=ccd_data)
instance = phase.constant
fit_figure_of_merit = analysis.fit(instance=instance)
mask = phase.mask_function(image=ccd_data.image)
lens_data = li.LensData(ccd_data=ccd_data, mask=mask)
tracer = analysis.tracer_for_instance(instance=instance)
fit = lens_fit.LensProfileFit(lens_data=lens_data, tracer=tracer)
assert fit.likelihood == fit_figure_of_merit
phase = ph.LensSourcePlanePhase(lens_galaxies=[lens_galaxy],
source_galaxies=[source_galaxy],
mask_function=ph.default_mask_function,
cosmology=cosmo.FLRW,
phase_name='test_phase')
analysis = phase.make_analysis(data=ccd_data)
instance = phase.constant
fit_figure_of_merit = analysis.fit(instance=instance)
mask = phase.mask_function(image=ccd_data.image)
lens_data = li.LensData(ccd_data=ccd_data, mask=mask)
tracer = analysis.tracer_for_instance(instance=instance)
fit = lens_fit.LensProfileInversionFit(lens_data=lens_data,
tracer=tracer)
assert fit.evidence == fit_figure_of_merit
def make_rectangular_pixelization():
return pix.Rectangular(shape=(25, 25))
def test___if_galaxy_has_pixelization__unmasked_image_is_none(self):
mask = msk.Mask(array=np.array([[True, True,
True], [True, False, True],
[True, True, True]]),
pixel_scale=1.0)
padded_grid_stack = grids.GridStack.padded_grid_stack_from_mask_sub_grid_size_and_psf_shape(
mask=mask, sub_grid_size=1, psf_shape=(3, 3))
psf_0 = im.PSF(array=(np.array([[0.0, 0.0, 0.0], [0.0, 1.0, 0.0],
[0.0, 0.0, 0.0]])),
pixel_scale=1.0)
psf_1 = im.PSF(array=(np.array([[0.0, 3.0, 0.0], [0.0, 1.0, 2.0],
[0.0, 0.0, 0.0]])),
pixel_scale=1.0)
g0 = g.Galaxy(light_profile=lp.EllipticalSersic(intensity=0.1),
pixelization=pix.Rectangular(),
regularization=reg.Constant())
g1 = g.Galaxy(light_profile=lp.EllipticalSersic(intensity=0.2))
g2 = g.Galaxy(mass_profile=mp.SphericalIsothermal(einstein_radius=1.0))
tracer = ray_tracing_stack.TracerImageSourcePlanesStack(
lens_galaxies=[g0, g1, g2],
source_galaxies=[g0, g1],
image_plane_grid_stacks=[padded_grid_stack, padded_grid_stack])
unmasked_blurred_image_of_datas_planes_and_galaxies = \
stack_util.unmasked_blurred_image_of_datas_planes_and_galaxies_from_padded_grid_stacks_and_psf(
planes=tracer.planes, padded_grid_stacks=[padded_grid_stack, padded_grid_stack], psfs=[psf_0, psf_1])
assert unmasked_blurred_image_of_datas_planes_and_galaxies[0][0][
0] == None
assert type(unmasked_blurred_image_of_datas_planes_and_galaxies[0][0]
[1]) == scaled_array.ScaledSquarePixelArray
assert unmasked_blurred_image_of_datas_planes_and_galaxies[1][0][
0] == None
assert type(unmasked_blurred_image_of_datas_planes_and_galaxies[1][0]
[1]) == scaled_array.ScaledSquarePixelArray
assert unmasked_blurred_image_of_datas_planes_and_galaxies[0][1][
0] == None
assert type(unmasked_blurred_image_of_datas_planes_and_galaxies[0][1]
[1]) == scaled_array.ScaledSquarePixelArray
assert unmasked_blurred_image_of_datas_planes_and_galaxies[1][1][
0] == None
assert type(unmasked_blurred_image_of_datas_planes_and_galaxies[1][1]
[1]) == scaled_array.ScaledSquarePixelArray
g0 = g.Galaxy(light_profile=lp.EllipticalSersic(intensity=0.1))
g1 = g.Galaxy(light_profile=lp.EllipticalSersic(intensity=0.2),
pixelization=pix.Rectangular(),
regularization=reg.Constant())
g2 = g.Galaxy(mass_profile=mp.SphericalIsothermal(einstein_radius=1.0))
tracer = ray_tracing_stack.TracerImageSourcePlanesStack(
lens_galaxies=[g0, g1, g2],
source_galaxies=[g0, g1],
image_plane_grid_stacks=[padded_grid_stack, padded_grid_stack])
unmasked_blurred_image_of_datas_planes_and_galaxies = \
stack_util.unmasked_blurred_image_of_datas_planes_and_galaxies_from_padded_grid_stacks_and_psf(
planes=tracer.planes, padded_grid_stacks=[padded_grid_stack, padded_grid_stack], psfs=[psf_0, psf_1])
assert type(unmasked_blurred_image_of_datas_planes_and_galaxies[0][0]
[0]) == scaled_array.ScaledSquarePixelArray
assert unmasked_blurred_image_of_datas_planes_and_galaxies[0][0][
1] == None
assert type(unmasked_blurred_image_of_datas_planes_and_galaxies[1][0]
[0]) == scaled_array.ScaledSquarePixelArray
assert unmasked_blurred_image_of_datas_planes_and_galaxies[1][0][
1] == None
assert type(unmasked_blurred_image_of_datas_planes_and_galaxies[0][1]
[0]) == scaled_array.ScaledSquarePixelArray
assert unmasked_blurred_image_of_datas_planes_and_galaxies[0][1][
1] == None
assert type(unmasked_blurred_image_of_datas_planes_and_galaxies[1][1]
[0]) == scaled_array.ScaledSquarePixelArray
assert unmasked_blurred_image_of_datas_planes_and_galaxies[1][1][
1] == None
def test__tracers_are_different__likelihood_is_non_zero(
self, lens_data_blur):
pixelization = pix.Rectangular(shape=(3, 3))
regularization = reg.Constant(coefficients=(1.0, ))
g0 = g.Galaxy(mass_profile=mp.SphericalIsothermal(einstein_radius=1.0))
g0_subhalo = g.Galaxy(subhalo=mp.SphericalIsothermal(
einstein_radius=0.1))
g1 = g.Galaxy(pixelization=pixelization, regularization=regularization)
tracer_normal = ray_tracing.TracerImageSourcePlanes(
lens_galaxies=[g0],
source_galaxies=[g1],
image_plane_grid_stack=lens_data_blur.grid_stack)
tracer_sensitive = ray_tracing.TracerImageSourcePlanes(
lens_galaxies=[g0, g0_subhalo],
source_galaxies=[g1],
image_plane_grid_stack=lens_data_blur.grid_stack)
fit = sensitivity_fit.SensitivityInversionFit(
lens_data=lens_data_blur,
tracer_normal=tracer_normal,
tracer_sensitive=tracer_sensitive)
assert (fit.fit_normal.image == lens_data_blur.image).all()
assert (fit.fit_normal.noise_map == lens_data_blur.noise_map).all()
mapper = pixelization.mapper_from_grid_stack_and_border(
grid_stack=tracer_normal.source_plane.grid_stack, border=None)
inversion = inv.inversion_from_image_mapper_and_regularization(
mapper=mapper,
regularization=regularization,
image_1d=lens_data_blur.image_1d,
noise_map_1d=lens_data_blur.noise_map_1d,
convolver=lens_data_blur.convolver_mapping_matrix)
assert fit.fit_normal.model_image == pytest.approx(
inversion.reconstructed_data, 1.0e-4)
residual_map = fit_util.residual_map_from_data_mask_and_model_data(
data=lens_data_blur.image,
mask=lens_data_blur.mask,
model_data=inversion.reconstructed_data)
assert fit.fit_normal.residual_map == pytest.approx(
residual_map, 1.0e-4)
chi_squared_map = fit_util.chi_squared_map_from_residual_map_noise_map_and_mask(
residual_map=residual_map,
mask=lens_data_blur.mask,
noise_map=lens_data_blur.noise_map)
assert fit.fit_normal.chi_squared_map == pytest.approx(
chi_squared_map, 1.0e-4)
assert (fit.fit_sensitive.image == lens_data_blur.image).all()
assert (fit.fit_sensitive.noise_map == lens_data_blur.noise_map).all()
mapper = pixelization.mapper_from_grid_stack_and_border(
grid_stack=tracer_sensitive.source_plane.grid_stack, border=None)
inversion = inv.inversion_from_image_mapper_and_regularization(
mapper=mapper,
regularization=regularization,
image_1d=lens_data_blur.image_1d,
noise_map_1d=lens_data_blur.noise_map_1d,
convolver=lens_data_blur.convolver_mapping_matrix)
assert fit.fit_sensitive.model_image == pytest.approx(
inversion.reconstructed_data, 1.0e-4)
residual_map = fit_util.residual_map_from_data_mask_and_model_data(
data=lens_data_blur.image,
mask=lens_data_blur.mask,
model_data=inversion.reconstructed_data)
assert fit.fit_sensitive.residual_map == pytest.approx(
residual_map, 1.0e-4)
chi_squared_map = fit_util.chi_squared_map_from_residual_map_noise_map_and_mask(
residual_map=residual_map,
mask=lens_data_blur.mask,
noise_map=lens_data_blur.noise_map)
assert fit.fit_sensitive.chi_squared_map == pytest.approx(
chi_squared_map, 1.0e-4)
chi_squared_normal = fit_util.chi_squared_from_chi_squared_map_and_mask(
chi_squared_map=fit.fit_normal.chi_squared_map,
mask=lens_data_blur.mask)
chi_squared_sensitive = fit_util.chi_squared_from_chi_squared_map_and_mask(
chi_squared_map=fit.fit_sensitive.chi_squared_map,
mask=lens_data_blur.mask)
noise_normalization = fit_util.noise_normalization_from_noise_map_and_mask(
mask=lens_data_blur.mask, noise_map=lens_data_blur.noise_map)
assert fit.fit_normal.likelihood == -0.5 * (chi_squared_normal +
noise_normalization)
assert fit.fit_sensitive.likelihood == -0.5 * (chi_squared_sensitive +
noise_normalization)
assert fit.figure_of_merit == fit.fit_sensitive.likelihood - fit.fit_normal.likelihood
fit_from_factory = sensitivity_fit.fit_lens_data_with_sensitivity_tracers(
lens_data=lens_data_blur,
tracer_normal=tracer_normal,
tracer_sensitive=tracer_sensitive)
assert fit.figure_of_merit == fit_from_factory.figure_of_merit
inner_radius_arcsec=1.0,
outer_radius_arcsec=2.2)
ccd_plotters.plot_image(ccd_data=ccd_data, mask=mask)
# Next, lets set this image up as lens data, and setup a tracer using the input lens galaxy model (we don't need
# to provide the source's light profile, as we're using a mapper to reconstruct it).
lens_data = ld.LensData(ccd_data=ccd_data, mask=mask, sub_grid_size=1)
lens_galaxy = g.Galaxy(mass=mp.EllipticalIsothermal(
centre=(0.0, 0.0), axis_ratio=0.8, phi=135.0, einstein_radius=1.6))
tracer = ray_tracing.TracerImageSourcePlanes(
lens_galaxies=[lens_galaxy],
source_galaxies=[g.Galaxy()],
image_plane_grid_stack=lens_data.grid_stack)
# We'll use another rectangular pixelization and mapper to perform the reconstruction
rectangular = pix.Rectangular(shape=(25, 25))
mapper = rectangular.mapper_from_grid_stack_and_border(
grid_stack=tracer.source_plane.grid_stack, border=None)
mapper_plotters.plot_image_and_mapper(ccd_data=ccd_data,
mask=mask,
mapper=mapper,
should_plot_grid=True)
# And now, finally, we're going to use our mapper to invert the image using the 'inversions' module, which is imported
# as 'inv'. I'll explain how this works in a second - but lets just go ahead and perform the inversion first.
# (Ignore the 'regularization' input below for now, we'll cover this in the next tutorial).
inversion = inv.Inversion(image_1d=lens_data.image_1d,
noise_map_1d=lens_data.noise_map_1d,
convolver=lens_data.convolver_mapping_matrix,
mapper=mapper,
regularization=reg.Constant(coefficients=(1.0, )))
shape=(100, 100), pixel_scale=0.05, sub_grid_size=2)
lens_galaxy = g.Galaxy(mass=mp.EllipticalIsothermal(
centre=(0.0, 0.0), axis_ratio=0.8, phi=90.0, einstein_radius=1.6))
# (Our source galaxy doesn't have a light profile from here on, as we're reconstructing its light using a pixelization).
tracer = ray_tracing.TracerImageSourcePlanes(
lens_galaxies=[lens_galaxy],
source_galaxies=[g.Galaxy()],
image_plane_grid_stack=image_plane_grid_stack)
# Next, lets set up a pixelization using the 'pixelizations' module, which we've imported as 'pix'.
# There are multiple pixelizations available in PyAutoLens, but for now we'll keep it simple and use a uniform
# rectangular grid. As usual, the grid's 'shape' defines its (y,x) dimensions.
rectangular = pix.Rectangular(shape=(25, 25))
# By itself, a pixelization doesn't tell us much. It has no grid of coordinates, no image, and nothing which tells it
# about the lens we're fitting. This information comes when we use the pixelization to set up a 'mapper'.
# (The 'border=None' will be covered in tutorial 5, so just ignore it for now!)
mapper = rectangular.mapper_from_grid_stack_and_border(
grid_stack=tracer.source_plane.grid_stack, border=None)
# This mapper is a 'RectangularMapper' - every pixelization generates it owns mapper.
print(type(mapper))
# By plotting our mapper, we now see our pixelization. Its a fairly boring grid of rectangular pixels.
mapper_plotters.plot_rectangular_mapper(
mapper=mapper,
should_plot_grid=False,
title='Fairly Boring Grid of Rectangular Pixels')
radius_arcsec=3.0)
ccd_plotters.plot_ccd_subplot(ccd_data=ccd_data, mask=mask)
# The lines of code below do everything we're used to, that is, setup an image and its grid stack, mask it, trace it
# via a tracer, setup the rectangular mapper, etc.
lens_galaxy = g.Galaxy(mass=mp.EllipticalIsothermal(
centre=(0.0, 0.0), axis_ratio=0.8, phi=45.0, einstein_radius=1.6))
lens_data = ld.LensData(ccd_data=ccd_data, mask=mask)
tracer = ray_tracing.TracerImageSourcePlanes(
lens_galaxies=[lens_galaxy],
source_galaxies=[g.Galaxy()],
image_plane_grid_stack=lens_data.grid_stack)
rectangular = pix.Rectangular(shape=(50, 50))
mapper = rectangular.mapper_from_grid_stack_and_border(
grid_stack=tracer.source_plane.grid_stack, border=None)
inversion = inv.Inversion(image_1d=lens_data.image_1d,
noise_map_1d=lens_data.noise_map_1d,
convolver=lens_data.convolver_mapping_matrix,
mapper=mapper,
regularization=reg.Constant(coefficients=(1.0, )))
# Now lets plot our rectangular mapper with the image.
mapper_plotters.plot_image_and_mapper(ccd_data=ccd_data,
mapper=mapper,
mask=mask,
should_plot_grid=True)
def test__number_of_pixels_and_regularization_set_up_correctly(self):
pix = pixelizations.Rectangular(shape=(3, 3))
assert pix.shape == (3, 3)
assert pix.pixels == 9
