def test__new_imaging_with_signal_to_noise_limit__include_mask_to_only_increase_centre_values(
self, ):
image = aa.Array2D.full(fill_value=20.0,
shape_native=(2, 2),
pixel_scales=1.0)
image[2] = 5.0
image[3] = 5.0
noise_map_array = aa.Array2D.full(fill_value=5.0,
shape_native=(2, 2),
pixel_scales=1.0)
noise_map_array[2] = 2.0
noise_map_array[3] = 2.0
mask = aa.Mask2D.manual(mask=[[True, False], [False, True]],
pixel_scales=1.0)
imaging = aa.Imaging(
image=image,
psf=aa.Kernel2D.zeros(shape_native=(3, 3), pixel_scales=1.0),
noise_map=noise_map_array,
)
imaging_capped = imaging.signal_to_noise_limited_from(
signal_to_noise_limit=2.0, mask=mask)
assert (imaging_capped.image.native == np.array([[20.0, 20.0],
[5.0, 5.0]])).all()
assert (imaging_capped.noise_map.native == np.array([[5.0, 10.0],
[2.5,
2.0]])).all()
assert (imaging_capped.signal_to_noise_map.native == np.array(
[[4.0, 2.0], [2.0, 2.5]])).all()
def test__new_imaging_with_signal_to_noise_limit__limit_above_max_signal_to_noise__signal_to_noise_map_unchanged(
self, ):
image = aa.Array2D.full(fill_value=20.0,
shape_native=(2, 2),
pixel_scales=1.0)
image[3] = 5.0
noise_map_array = aa.Array2D.full(fill_value=5.0,
shape_native=(2, 2),
pixel_scales=1.0)
noise_map_array[3] = 2.0
imaging = aa.Imaging(
image=image,
psf=aa.Kernel2D.zeros(shape_native=(3, 3), pixel_scales=1.0),
noise_map=noise_map_array,
)
imaging = imaging.signal_to_noise_limited_from(
signal_to_noise_limit=100.0)
assert (imaging.image == np.array([20.0, 20.0, 20.0, 5.0])).all()
assert (imaging.noise_map == np.array([5.0, 5.0, 5.0, 2.0])).all()
assert (imaging.signal_to_noise_map == np.array([4.0, 4.0, 4.0,
2.5])).all()
def test__new_imaging_with_signal_to_noise_limit_below_max_signal_to_noise__signal_to_noise_map_capped_to_limit(
self, ):
image = aa.Array2D.full(fill_value=20.0,
shape_native=(2, 2),
pixel_scales=1.0)
image[3] = 5.0
noise_map_array = aa.Array2D.full(fill_value=5.0,
shape_native=(2, 2),
pixel_scales=1.0)
noise_map_array[3] = 2.0
imaging = aa.Imaging(
image=image,
psf=aa.Kernel2D.zeros(shape_native=(3, 3), pixel_scales=1.0),
noise_map=noise_map_array,
)
imaging_capped = imaging.signal_to_noise_limited_from(
signal_to_noise_limit=2.0)
assert (imaging_capped.image.native == np.array([[20.0, 20.0],
[20.0, 5.0]])).all()
assert (imaging_capped.noise_map.native == np.array([[10.0, 10.0],
[10.0,
2.5]])).all()
assert (imaging_capped.signal_to_noise_map.native == np.array(
[[2.0, 2.0], [2.0, 2.0]])).all()
def test__different_imaging_without_mock_objects__customize_constructor_inputs(
self, ):
psf = aa.Kernel2D.ones(shape_native=(7, 7), pixel_scales=3.0)
imaging = aa.Imaging(
image=aa.Array2D.ones(shape_native=(19, 19), pixel_scales=3.0),
psf=psf,
noise_map=aa.Array2D.full(fill_value=2.0,
shape_native=(19, 19),
pixel_scales=3.0),
)
mask = aa.Mask2D.unmasked(shape_native=(19, 19),
pixel_scales=1.0,
invert=True,
sub_size=8)
mask[9, 9] = False
masked_imaging = imaging.apply_mask(mask=mask)
assert masked_imaging.psf.native == pytest.approx(
(1.0 / 49.0) * np.ones((7, 7)), 1.0e-4)
assert masked_imaging.convolver.kernel.shape_native == (7, 7)
assert (masked_imaging.image == np.array([1.0])).all()
assert (masked_imaging.noise_map == np.array([2.0])).all()
def _imaging_from(fit: af.Fit, settings_imaging: Optional[aa.SettingsImaging] = None):
"""
Returns a `Imaging` object from an aggregator's `SearchOutput` class, which we call an 'agg_obj' to describe
that it acts as the aggregator object for one result in the `Aggregator`. This uses the aggregator's generator
outputs such that the function can use the `Aggregator`'s map function to to create a `Imaging` generator.
The `Imaging` is created following the same method as the PyAutoGalaxy `Search` classes, including using the
`SettingsImaging` instance output by the Search to load inputs of the `Imaging` (e.g. psf_shape_2d).
Parameters
----------
fit : ImaginSearchOutput
A PyAutoFit aggregator's SearchOutput object containing the generators of the results of PyAutoGalaxy model-fits.
"""
data = fit.value(name="data")
noise_map = fit.value(name="noise_map")
psf = fit.value(name="psf")
settings_imaging = settings_imaging or fit.value(name="settings_dataset")
if not hasattr(settings_imaging, "relative_accuracy"):
settings_imaging.relative_accuracy = None
imaging = aa.Imaging(
image=data,
noise_map=noise_map,
psf=psf,
settings=settings_imaging,
pad_for_convolver=True,
)
imaging.apply_settings(settings=settings_imaging)
return imaging
def test__image_and_model_are_different__include_masking__check_values_are_correct(
self, ):
mask = aa.Mask2D.manual(mask=[[False, False], [True, False]],
sub_size=1,
pixel_scales=(1.0, 1.0))
data = aa.Array2D.manual_mask(array=[1.0, 2.0, 4.0], mask=mask)
noise_map = aa.Array2D.manual_mask(array=[2.0, 2.0, 2.0], mask=mask)
imaging = aa.Imaging(image=data, noise_map=noise_map)
model_image = aa.Array2D.manual_mask(array=[1.0, 2.0, 3.0], mask=mask)
fit = aa.FitImaging(imaging=imaging,
model_image=model_image,
use_mask_in_fit=False)
assert (fit.mask == np.array([[False, False], [True, False]])).all()
assert (fit.image.slim == np.array([1.0, 2.0, 4.0])).all()
assert (fit.image.native == np.array([[1.0, 2.0], [0.0, 4.0]])).all()
assert (fit.noise_map.slim == np.array([2.0, 2.0, 2.0])).all()
assert (fit.noise_map.native == np.array([[2.0, 2.0], [0.0,
2.0]])).all()
assert (fit.signal_to_noise_map.slim == np.array([0.5, 1.0,
2.0])).all()
assert (fit.signal_to_noise_map.native == np.array([[0.5, 1.0],
[0.0,
2.0]])).all()
assert (fit.model_image.slim == np.array([1.0, 2.0, 3.0])).all()
assert (fit.model_image.native == np.array([[1.0, 2.0], [0.0,
3.0]])).all()
assert (fit.residual_map.slim == np.array([0.0, 0.0, 1.0])).all()
assert (fit.residual_map.native == np.array([[0.0, 0.0],
[0.0, 1.0]])).all()
assert (fit.normalized_residual_map.slim == np.array([0.0, 0.0,
0.5])).all()
assert (fit.normalized_residual_map.native == np.array([[0.0, 0.0],
[0.0,
0.5]])).all()
assert (fit.chi_squared_map.slim == np.array([0.0, 0.0, 0.25])).all()
assert (fit.chi_squared_map.native == np.array([[0.0, 0.0],
[0.0, 0.25]])).all()
assert fit.chi_squared == 0.25
assert fit.reduced_chi_squared == 0.25 / 3.0
assert fit.noise_normalization == np.sum(
np.log(2 * np.pi * noise_map**2.0))
assert fit.log_likelihood == -0.5 * (fit.chi_squared +
fit.noise_normalization)
def test__psf_and_mask_hit_edge__automatically_pads_image_and_noise_map(
self):
image = aa.Array2D.ones(shape_native=(3, 3), pixel_scales=1.0)
noise_map = aa.Array2D.ones(shape_native=(3, 3), pixel_scales=1.0)
psf = aa.Kernel2D.ones(shape_native=(3, 3), pixel_scales=1.0)
imaging = aa.Imaging(image=image,
noise_map=noise_map,
psf=psf,
setup_convolver=False)
assert imaging.image.shape_native == (3, 3)
assert imaging.noise_map.shape_native == (3, 3)
imaging = aa.Imaging(image=image,
noise_map=noise_map,
psf=psf,
setup_convolver=True)
assert imaging.image.shape_native == (5, 5)
assert imaging.noise_map.shape_native == (5, 5)
assert imaging.image.mask[0, 0] == True
assert imaging.image.mask[1, 1] == False
def test__image_and_model_are_identical__inversion_included__changes_certain_properties(
self, ):
mask = aa.Mask2D.manual(mask=[[False, False], [False, False]],
sub_size=1,
pixel_scales=(1.0, 1.0))
data = aa.Array2D.manual_mask(array=[1.0, 2.0, 3.0, 4.0], mask=mask)
noise_map = aa.Array2D.manual_mask(array=[2.0, 2.0, 2.0, 2.0],
mask=mask)
imaging = aa.Imaging(image=data, noise_map=noise_map)
masked_imaging = imaging.apply_mask(mask=mask)
model_image = aa.Array2D.manual_mask(array=[1.0, 2.0, 3.0, 4.0],
mask=mask)
inversion = mock.MockFitInversion(
regularization_term=2.0,
log_det_curvature_reg_matrix_term=3.0,
log_det_regularization_matrix_term=4.0,
)
fit = aa.FitImaging(
imaging=masked_imaging,
model_image=model_image,
inversion=inversion,
use_mask_in_fit=False,
)
assert fit.chi_squared == 0.0
assert fit.reduced_chi_squared == 0.0
assert fit.noise_normalization == np.sum(
np.log(2 * np.pi * noise_map**2.0))
assert fit.log_likelihood == -0.5 * (fit.chi_squared +
fit.noise_normalization)
assert fit.log_likelihood_with_regularization == -0.5 * (
fit.chi_squared + 2.0 + fit.noise_normalization)
assert fit.log_evidence == -0.5 * (fit.chi_squared + 2.0 + 3.0 - 4.0 +
fit.noise_normalization)
assert fit.figure_of_merit == fit.log_evidence
def test__15_grid__no_sub_grid(self):
mask = aa.Mask2D.manual(
mask=[
[True, True, True, True, True, True, True],
[True, True, True, True, True, True, True],
[True, False, False, False, False, False, True],
[True, False, False, False, False, False, True],
[True, False, False, False, False, False, True],
[True, True, True, True, True, True, True],
[True, True, True, True, True, True, True],
],
pixel_scales=1.0,
sub_size=1,
)
# There is no sub-grid, so our grid are just the masked_image grid (note the NumPy weighted_data structure
# ensures this has no sub-gridding)
grid = aa.Grid2D.manual_mask(
grid=[
[0.9, -0.9],
[1.0, -1.0],
[1.1, -1.1],
[0.9, 0.9],
[1.0, 1.0],
[1.1, 1.1],
[-0.01, 0.01],
[0.0, 0.0],
[0.01, 0.01],
[-0.9, -0.9],
[-1.0, -1.0],
[-1.1, -1.1],
[-0.9, 0.9],
[-1.0, 1.0],
[-1.1, 1.1],
],
mask=mask,
)
pix = aa.pix.Rectangular(shape=(3, 3))
mapper = pix.mapper_from_grid_and_sparse_grid(
grid=grid,
sparse_grid=None,
settings=aa.SettingsPixelization(use_border=False),
)
assert mapper.data_pixelization_grid == None
assert mapper.source_pixelization_grid.shape_native_scaled == pytest.approx(
(2.2, 2.2), 1.0e-4)
assert mapper.source_pixelization_grid.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],
[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, 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, 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, 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, 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],
])).all()
assert mapper.shape_native == (3, 3)
reg = aa.reg.Constant(coefficient=1.0)
regularization_matrix = reg.regularization_matrix_from_mapper(
mapper=mapper)
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()
image = aa.Array2D.ones(shape_native=(7, 7), pixel_scales=1.0)
noise_map = aa.Array2D.ones(shape_native=(7, 7), pixel_scales=1.0)
psf = aa.Kernel2D.no_blur(pixel_scales=1.0)
imaging = aa.Imaging(image=image, noise_map=noise_map, psf=psf)
masked_imaging = imaging.apply_mask(mask=mask)
inversion = aa.Inversion(
dataset=masked_imaging,
mapper=mapper,
regularization=reg,
settings=aa.SettingsInversion(check_solution=False),
)
assert (
inversion.blurred_mapping_matrix == mapper.mapping_matrix).all()
assert (inversion.regularization_matrix == regularization_matrix).all()
assert inversion.mapped_reconstructed_image == pytest.approx(
np.ones(15), 1.0e-4)
def test__3x3_simple_grid__include_mask_with_offset_centre(self):
mask = aa.Mask2D.manual(
mask=[
[True, True, True, True, True, True, True],
[True, True, True, True, False, True, True],
[True, True, True, False, False, False, True],
[True, True, True, True, False, True, True],
[True, True, True, True, True, True, True],
[True, True, True, True, True, True, True],
[True, True, True, True, True, True, True],
],
pixel_scales=1.0,
sub_size=1,
)
grid = np.array([[2.0, 1.0], [1.0, 0.0], [1.0, 1.0], [1.0, 2.0],
[0.0, 1.0]])
grid = aa.Grid2D.manual_mask(grid=grid, mask=mask)
pix = aa.pix.VoronoiMagnification(shape=(3, 3))
sparse_grid = aa.Grid2DSparse.from_grid_and_unmasked_2d_grid_shape(
grid=grid, unmasked_sparse_shape=pix.shape)
mapper = pix.mapper_from_grid_and_sparse_grid(
grid=grid,
sparse_grid=sparse_grid,
settings=aa.SettingsPixelization(use_border=False),
)
assert mapper.source_pixelization_grid.shape_native_scaled == pytest.approx(
(2.0, 2.0), 1.0e-4)
assert (mapper.source_pixelization_grid == sparse_grid).all()
# assert mapper.pixelization_grid.origin == pytest.approx((1.0, 1.0), 1.0e-4)
assert isinstance(mapper, mappers.MapperVoronoi)
assert (mapper.mapping_matrix == np.array([
[1.0, 0.0, 0.0, 0.0, 0.0],
[0.0, 1.0, 0.0, 0.0, 0.0],
[0.0, 0.0, 1.0, 0.0, 0.0],
[0.0, 0.0, 0.0, 1.0, 0.0],
[0.0, 0.0, 0.0, 0.0, 1.0],
])).all()
reg = aa.reg.Constant(coefficient=1.0)
regularization_matrix = reg.regularization_matrix_from_mapper(
mapper=mapper)
assert (regularization_matrix == np.array([
[3.00000001, -1.0, -1.0, -1.0, 0.0],
[-1.0, 3.00000001, -1.0, 0.0, -1.0],
[-1.0, -1.0, 4.00000001, -1.0, -1.0],
[-1.0, 0.0, -1.0, 3.00000001, -1.0],
[0.0, -1.0, -1.0, -1.0, 3.00000001],
])).all()
image = aa.Array2D.ones(shape_native=(7, 7), pixel_scales=1.0)
noise_map = aa.Array2D.ones(shape_native=(7, 7), pixel_scales=1.0)
psf = aa.Kernel2D.no_blur(pixel_scales=1.0)
imaging = aa.Imaging(image=image, noise_map=noise_map, psf=psf)
masked_imaging = imaging.apply_mask(mask=mask)
inversion = aa.Inversion(
dataset=masked_imaging,
mapper=mapper,
regularization=reg,
settings=aa.SettingsInversion(check_solution=False),
)
assert (
inversion.blurred_mapping_matrix == mapper.mapping_matrix).all()
assert (inversion.regularization_matrix == regularization_matrix).all()
assert inversion.mapped_reconstructed_image == pytest.approx(
np.ones(5), 1.0e-4)
def test__5_simple_grid__include_sub_grid(self):
mask = aa.Mask2D.manual(
mask=[
[True, True, True, True, True, True, True],
[True, True, True, True, True, True, True],
[True, True, True, False, True, True, True],
[True, True, False, False, False, True, True],
[True, True, True, False, True, True, True],
[True, True, True, True, True, True, True],
[True, True, True, True, True, True, True],
],
pixel_scales=2.0,
sub_size=2,
)
# 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 lens calculation. This is to make a mapping_matrix matrix which explicitly tests the
# sub-grid.
grid = aa.Grid2D.manual_mask(
grid=[
[1.0, -1.0],
[1.0, -1.0],
[1.0, -1.0],
[1.0, 1.0],
[1.0, 1.0],
[1.0, 1.0],
[-1.0, -1.0],
[-1.0, -1.0],
[-1.0, -1.0],
[-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],
[0.0, 0.0],
[0.0, 0.0],
[0.0, 0.0],
],
mask=mask,
)
pix = aa.pix.Rectangular(shape=(3, 3))
mapper = pix.mapper_from_grid_and_sparse_grid(
grid=grid,
sparse_grid=None,
settings=aa.SettingsPixelization(use_border=False),
)
assert mapper.data_pixelization_grid == None
assert mapper.source_pixelization_grid.shape_native_scaled == pytest.approx(
(2.0, 2.0), 1.0e-4)
assert mapper.source_pixelization_grid.origin == pytest.approx(
(0.0, 0.0), 1.0e-4)
assert (mapper.mapping_matrix == np.array([
[0.75, 0.0, 0.25, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0],
[0.0, 0.0, 0.5, 0.0, 0.0, 0.0, 0.5, 0.0, 0.0],
[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.25, 0.0, 0.75],
[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],
])).all()
assert mapper.shape_native == (3, 3)
reg = aa.reg.Constant(coefficient=1.0)
regularization_matrix = reg.regularization_matrix_from_mapper(
mapper=mapper)
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()
image = aa.Array2D.ones(shape_native=(7, 7), pixel_scales=1.0)
noise_map = aa.Array2D.ones(shape_native=(7, 7), pixel_scales=1.0)
psf = aa.Kernel2D.no_blur(pixel_scales=1.0)
imaging = aa.Imaging(image=image, noise_map=noise_map, psf=psf)
masked_imaging = imaging.apply_mask(mask=mask)
inversion = aa.Inversion(
dataset=masked_imaging,
mapper=mapper,
regularization=reg,
settings=aa.SettingsInversion(check_solution=False),
)
assert (
inversion.blurred_mapping_matrix == mapper.mapping_matrix).all()
assert (inversion.regularization_matrix == regularization_matrix).all()
assert inversion.mapped_reconstructed_image == pytest.approx(
np.ones(5), 1.0e-4)
