refactor dataset and operator tests for granularity and clarity

Split monolithic tests into focused single-assertion tests with
descriptive names encoding condition and expected result. Add
@pytest.mark.parametrize for convolver and transformer scenarios.
Also fix pytest assert-tuple bug in convolver test.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Author: 2 people

test_autoarray/dataset/imaging/test_dataset.py

@@ -33,7 +33,7 @@ def make_test_data_path():
     return test_data_path
 
 
-def test__grid__uses_mask_and_settings(
+def test__grid__uses_mask_and_settings__lp_grid_matches_grid_2d_7x7(
     image_7x7,
     noise_map_7x7,
     mask_2d_7x7,
@@ -57,7 +57,7 @@ def test__grid__uses_mask_and_settings(
     assert (masked_imaging_7x7.grids.lp.slim == grid_2d_7x7).all()
 
 
-def test__grids_pixelization__uses_mask_and_settings(
+def test__grids_pixelization__uses_mask_and_settings__default_over_sample__matches_grid_2d_7x7(
     image_7x7,
     noise_map_7x7,
     mask_2d_7x7,
@@ -75,6 +75,17 @@ def test__grids_pixelization__uses_mask_and_settings(
     assert (masked_imaging_7x7.grids.pixelization == grid_2d_7x7).all()
     assert (masked_imaging_7x7.grids.pixelization.slim == grid_2d_7x7).all()
 
+
+def test__grids_pixelization__uses_mask_and_settings__custom_over_sample__returns_grid2d_with_correct_size(
+    image_7x7,
+    noise_map_7x7,
+    mask_2d_7x7,
+    grid_2d_7x7,
+):
+    masked_image_7x7 = aa.Array2D(values=image_7x7.native, mask=mask_2d_7x7)
+
+    masked_noise_map_7x7 = aa.Array2D(values=noise_map_7x7.native, mask=mask_2d_7x7)
+
     masked_imaging_7x7 = aa.Imaging(
         data=masked_image_7x7,
         noise_map=masked_noise_map_7x7,
@@ -86,7 +97,9 @@ def test__grids_pixelization__uses_mask_and_settings(
     assert masked_imaging_7x7.grids.over_sample_size_pixelization[0] == 4
 
 
-def test__grid_settings__sub_size(image_7x7, noise_map_7x7):
+def test__grid_settings__sub_size__returns_correct_over_sample_sizes(
+    image_7x7, noise_map_7x7
+):
     dataset_7x7 = aa.Imaging(
         data=image_7x7,
         noise_map=noise_map_7x7,
@@ -116,7 +129,7 @@ def test__no_noise_map__raises_exception():
         aa.Imaging(data=image)
 
 
-def test__from_fits():
+def test__from_fits__separate_fits_files__loads_data_psf_noise_map_correctly():
     dataset = aa.Imaging.from_fits(
         pixel_scales=0.1,
         data_path=path.join(test_data_path, "3x3_ones.fits"),
@@ -134,6 +147,8 @@ def test__from_fits():
     assert dataset.psf.kernel.mask.pixel_scales == (0.1, 0.1)
     assert dataset.noise_map.mask.pixel_scales == (0.1, 0.1)
 
+
+def test__from_fits__all_data_in_one_fits_file_multiple_hdus__loads_data_psf_noise_map_correctly():
     dataset = aa.Imaging.from_fits(
         pixel_scales=0.1,
         data_path=path.join(test_data_path, "3x3_multiple_hdu.fits"),
@@ -155,7 +170,9 @@ def test__from_fits():
     assert dataset.noise_map.mask.pixel_scales == (0.1, 0.1)
 
 
-def test__output_to_fits(imaging_7x7, test_data_path):
+def test__output_to_fits__round_trips_data_psf_noise_map_correctly(
+    imaging_7x7, test_data_path
+):
 
     imaging_7x7.output_to_fits(
         data_path=path.join(test_data_path, "data.fits"),
@@ -176,7 +193,9 @@ def test__output_to_fits(imaging_7x7, test_data_path):
     assert dataset.pixel_scales == (0.1, 0.1)
 
 
-def test__apply_mask(imaging_7x7, mask_2d_7x7, psf_3x3):
+def test__apply_mask__data_noise_map_psf_correctly_masked(
+    imaging_7x7, mask_2d_7x7, psf_3x3
+):
     masked_imaging_7x7 = imaging_7x7.apply_mask(mask=mask_2d_7x7)
 
     assert (masked_imaging_7x7.data.slim == np.ones(9)).all()
@@ -198,7 +217,9 @@ def test__apply_mask(imaging_7x7, mask_2d_7x7, psf_3x3):
     assert type(masked_imaging_7x7.psf) == aa.Convolver
 
 
-def test__apply_noise_scaling(imaging_7x7, mask_2d_7x7):
+def test__apply_noise_scaling__masked_pixel_data_zeroed_and_noise_set_to_noise_value(
+    imaging_7x7, mask_2d_7x7
+):
     masked_imaging_7x7 = imaging_7x7.apply_noise_scaling(
         mask=mask_2d_7x7, noise_value=1e5
     )
@@ -207,7 +228,7 @@ def test__apply_noise_scaling(imaging_7x7, mask_2d_7x7):
     assert masked_imaging_7x7.noise_map.native[4, 4] == 1e5
 
 
-def test__apply_noise_scaling__use_signal_to_noise_value(
+def test__apply_noise_scaling__use_signal_to_noise_value__noise_map_scaled_to_match_snr(
     image_7x7, psf_3x3, noise_map_7x7, mask_2d_7x7
 ):
 
@@ -233,7 +254,7 @@ def test__apply_noise_scaling__use_signal_to_noise_value(
     assert masked_imaging_7x7.noise_map.native[3, 3] == 10.0
 
 
-def test__apply_mask__noise_covariance_matrix():
+def test__apply_mask__noise_covariance_matrix__submatrix_extracted_for_unmasked_pixels():
     image = aa.Array2D.ones(shape_native=(2, 2), pixel_scales=(1.0, 1.0))
 
     noise_covariance_matrix = np.array(
@@ -263,7 +284,7 @@ def test__apply_mask__noise_covariance_matrix():
     )
 
 
-def test__different_imaging_without_mock_objects__customize_constructor_inputs():
+def test__different_imaging_without_mock_objects__customize_constructor_inputs__single_unmasked_pixel_correct():
 
     kernel = aa.Array2D.ones(shape_native=(7, 7), pixel_scales=3.0)
     psf = aa.Convolver(kernel=kernel)
@@ -291,14 +312,18 @@ def test__different_imaging_without_mock_objects__customize_constructor_inputs()
     assert (masked_dataset.noise_map == np.array([2.0])).all()
 
 
-def test__noise_map_unmasked_has_zeros_or_negative__raises_exception():
+def test__noise_map_unmasked_has_zeros__raises_exception():
     array = aa.Array2D.no_mask([[1.0, 2.0]], pixel_scales=1.0)
 
     noise_map = aa.Array2D.no_mask([[0.0, 3.0]], pixel_scales=1.0)
 
     with pytest.raises(aa.exc.DatasetException):
         aa.Imaging(data=array, noise_map=noise_map)
 
+
+def test__noise_map_unmasked_has_negative_values__raises_exception():
+    array = aa.Array2D.no_mask([[1.0, 2.0]], pixel_scales=1.0)
+
     noise_map = aa.Array2D.no_mask([[-1.0, 3.0]], pixel_scales=1.0)
 
     with pytest.raises(aa.exc.DatasetException):

test_autoarray/dataset/interferometer/test_dataset.py

@@ -14,7 +14,7 @@
 )
 
 
-def test__dirty_properties(
+def test__dirty_image__shape_native_matches_real_space_mask(
     visibilities_7,
     visibilities_noise_map_7,
     uv_wavelengths_7x2,
@@ -28,12 +28,39 @@ def test__dirty_properties(
     )
 
     assert dataset.dirty_image.shape_native == (7, 7)
-
     assert (dataset.transformer.image_from(visibilities=dataset.data)).all()
 
+
+def test__dirty_noise_map__shape_native_matches_real_space_mask(
+    visibilities_7,
+    visibilities_noise_map_7,
+    uv_wavelengths_7x2,
+    mask_2d_7x7,
+):
+    dataset = aa.Interferometer(
+        data=visibilities_7,
+        noise_map=visibilities_noise_map_7,
+        uv_wavelengths=uv_wavelengths_7x2,
+        real_space_mask=mask_2d_7x7,
+    )
+
     assert dataset.dirty_noise_map.shape_native == (7, 7)
     assert (dataset.transformer.image_from(visibilities=dataset.noise_map)).all()
 
+
+def test__dirty_signal_to_noise_map__shape_native_matches_real_space_mask(
+    visibilities_7,
+    visibilities_noise_map_7,
+    uv_wavelengths_7x2,
+    mask_2d_7x7,
+):
+    dataset = aa.Interferometer(
+        data=visibilities_7,
+        noise_map=visibilities_noise_map_7,
+        uv_wavelengths=uv_wavelengths_7x2,
+        real_space_mask=mask_2d_7x7,
+    )
+
     assert dataset.dirty_signal_to_noise_map.shape_native == (7, 7)
     assert (
         dataset.transformer.image_from(visibilities=dataset.signal_to_noise_map)
@@ -102,7 +129,7 @@ def test__output_all_arrays(mask_2d_7x7):
     assert (dataset.uv_wavelengths[:, 1] == 6.0 * np.ones(3)).all()
 
 
-def test__transformer(
+def test__transformer__dft_class__returns_transformer_dft_instance(
     visibilities_7,
     visibilities_noise_map_7,
     uv_wavelengths_7x2,
@@ -118,6 +145,13 @@ def test__transformer(
 
     assert type(interferometer_7.transformer) == transformer.TransformerDFT
 
+
+def test__transformer__nufft_class__returns_transformer_nufft_instance(
+    visibilities_7,
+    visibilities_noise_map_7,
+    uv_wavelengths_7x2,
+    mask_2d_7x7,
+):
     interferometer_7 = aa.Interferometer(
         data=visibilities_7,
         noise_map=visibilities_noise_map_7,

test_autoarray/mask/derive/test_mask_2d.py

@@ -24,24 +24,29 @@ def make_derive_mask_2d_9x9():
     return aa.DeriveMask2D(mask=mask_2d)
 
 
-def test__unmasked_mask(derive_mask_2d_9x9):
+def test__unmasked_mask__all_false__returns_full_false_array(derive_mask_2d_9x9):
     assert (
         derive_mask_2d_9x9.all_false == np.full(fill_value=False, shape=(9, 9))
     ).all()
 
 
-def test__blurring_mask_from(derive_mask_2d_9x9):
+@pytest.mark.parametrize("kernel_shape_native", [(3, 3)])
+def test__blurring_mask_from__3x3_kernel__matches_util_result(
+    derive_mask_2d_9x9, kernel_shape_native
+):
     blurring_mask_via_util = aa.util.mask_2d.blurring_mask_2d_from(
         mask_2d=derive_mask_2d_9x9.mask,
-        kernel_shape_native=(3, 3),
+        kernel_shape_native=kernel_shape_native,
     )
 
-    blurring_mask = derive_mask_2d_9x9.blurring_from(kernel_shape_native=(3, 3))
+    blurring_mask = derive_mask_2d_9x9.blurring_from(
+        kernel_shape_native=kernel_shape_native
+    )
 
     assert (blurring_mask == blurring_mask_via_util).all()
 
 
-def test__edge_mask(derive_mask_2d_9x9):
+def test__edge_mask__9x9_ring_mask__edge_pixels_are_unmasked(derive_mask_2d_9x9):
     assert (
         derive_mask_2d_9x9.edge
         == np.array(
@@ -60,7 +65,7 @@ def test__edge_mask(derive_mask_2d_9x9):
     ).all()
 
 
-def test__edge_buffed_mask():
+def test__edge_buffed_mask__5x5_mask_with_centre_masked__buffed_mask_matches_util():
     mask = aa.Mask2D.all_false(shape_native=(5, 5), pixel_scales=1.0)
     mask[2, 2] = True
 
@@ -73,7 +78,7 @@ def test__edge_buffed_mask():
     assert (derive_mask_2d.edge_buffed == edge_buffed_mask_manual).all()
 
 
-def test__border_mask(derive_mask_2d_9x9):
+def test__border_mask__9x9_ring_mask__inner_pixels_are_masked(derive_mask_2d_9x9):
     assert (
         derive_mask_2d_9x9.border
         == np.array(