class ImgDataUnitTest(unittest.TestCase):
def setUp(self):
self.app = QtGui.QApplication([])
self.img_model = ImgModel()
self.img_model.load(os.path.join(data_path, 'image_001.tif'))
def tearDown(self):
del self.app
del self.img_model
gc.collect()
def perform_transformations_tests(self):
self.assertEqual(np.sum(np.absolute(self.img_model.get_img_data())), 0)
self.img_model.rotate_img_m90()
self.assertEqual(np.sum(np.absolute(self.img_model.get_img_data())), 0)
self.img_model.flip_img_horizontally()
self.assertEqual(np.sum(np.absolute(self.img_model.get_img_data())), 0)
self.img_model.rotate_img_p90()
self.assertEqual(np.sum(np.absolute(self.img_model.get_img_data())), 0)
self.img_model.flip_img_vertically()
self.assertEqual(np.sum(np.absolute(self.img_model.get_img_data())), 0)
self.img_model.reset_img_transformations()
self.assertEqual(np.sum(np.absolute(self.img_model.get_img_data())), 0)
def test_flipping_images(self):
original_image = np.copy(self.img_model._img_data)
self.img_model.flip_img_vertically()
self.assertTrue(np.array_equal(self.img_model._img_data, np.flipud(original_image)))
def test_simple_background_subtraction(self):
self.first_image = np.copy(self.img_model.get_img_data())
self.img_model.load_next_file()
self.second_image = np.copy(self.img_model.get_img_data())
self.img_model.load(os.path.join(data_path, 'image_001.tif'))
self.img_model.load_background(os.path.join(data_path, 'image_002.tif'))
self.assertFalse(np.array_equal(self.first_image, self.img_model.get_img_data()))
self.img_model.load_next_file()
self.assertEqual(np.sum(self.img_model.get_img_data()), 0)
def test_background_subtraction_with_supersampling(self):
self.img_model.load_background(os.path.join(data_path, 'image_002.tif'))
self.img_model.set_supersampling(2)
self.img_model.get_img_data()
self.img_model.set_supersampling(3)
self.img_model.get_img_data()
self.img_model.load_next_file()
self.img_model.get_img_data()
def test_background_subtraction_with_transformation(self):
self.img_model.load_background(os.path.join(data_path, 'image_002.tif'))
original_img = np.copy(self.img_model._img_data)
original_background = np.copy(self.img_model._background_data)
self.assertNotEqual(self.img_model._background_data, None)
self.assertFalse(np.array_equal(self.img_model.img_data, self.img_model._img_data))
original_img_background_subtracted = np.copy(self.img_model.get_img_data())
self.assertTrue(np.array_equal(original_img_background_subtracted, original_img-original_background))
### now comes the main process - flipping the image
self.img_model.flip_img_vertically()
flipped_img = np.copy(self.img_model._img_data)
self.assertTrue(np.array_equal(np.flipud(original_img), flipped_img))
flipped_background = np.copy(self.img_model._background_data)
self.assertTrue(np.array_equal(np.flipud(original_background), flipped_background))
flipped_img_background_subtracted = np.copy(self.img_model.get_img_data())
self.assertTrue(np.array_equal(flipped_img_background_subtracted, flipped_img-flipped_background))
self.assertTrue(np.array_equal(np.flipud(original_img_background_subtracted),
flipped_img_background_subtracted))
self.assertEqual(np.sum(np.flipud(original_img_background_subtracted)-flipped_img_background_subtracted), 0)
self.img_model.load(os.path.join(data_path, 'image_002.tif'))
self.perform_transformations_tests()
def test_background_subtraction_with_supersampling_and_image_transformation(self):
self.img_model.load_background(os.path.join(data_path, 'image_002.tif'))
self.img_model.load(os.path.join(data_path, 'image_002.tif'))
self.img_model.set_supersampling(2)
self.assertEqual(self.img_model.get_img_data().shape, (4096, 4096))
self.perform_transformations_tests()
self.img_model.set_supersampling(3)
self.assertEqual(self.img_model.get_img_data().shape, (6144, 6144))
self.perform_transformations_tests()
self.img_model.load(os.path.join(data_path, 'image_002.tif'))
self.assertEqual(self.img_model.get_img_data().shape, (6144, 6144))
self.perform_transformations_tests()
def test_background_scaling_and_offset(self):
self.img_model.load_background(os.path.join(data_path, 'image_002.tif'))
#assure that everything is correct before
self.assertTrue(np.array_equal(self.img_model.get_img_data(),
self.img_model._img_data-self.img_model._background_data))
#set scaling and see difference
self.img_model.set_background_scaling(2.4)
self.assertTrue(np.array_equal(self.img_model.get_img_data(),
self.img_model._img_data-2.4*self.img_model._background_data))
#set offset and see the difference
self.img_model.set_background_scaling(1.0)
self.img_model.set_background_offset(100.0)
self.assertTrue(np.array_equal(self.img_model.img_data,
self.img_model._img_data-(self.img_model._background_data+100.0)))
#use offset and scaling combined
self.img_model.set_background_scaling(2.3)
self.img_model.set_background_offset(100.0)
self.assertTrue(np.array_equal(self.img_model.img_data,
self.img_model._img_data-(2.3*self.img_model._background_data+100)))
def test_background_with_different_shape(self):
self.img_model.load_background(os.path.join(data_path, 'CeO2_Pilatus1M.tif'))
self.assertEqual(self.img_model._background_data, None)
self.img_model.load_background(os.path.join(data_path, 'image_002.tif'))
self.assertTrue(self.img_model._background_data is not None)
self.img_model.load(os.path.join(data_path, 'CeO2_Pilatus1M.tif'))
self.assertEqual(self.img_model._background_data, None)
def test_absorption_correction_with_supersampling(self):
original_image = np.copy(self.img_model.get_img_data())
dummy_correction = DummyCorrection(self.img_model.get_img_data().shape, 0.6)
self.img_model.add_img_correction(dummy_correction, "Dummy 1")
self.assertAlmostEqual(np.sum(original_image)/0.6, np.sum(self.img_model.get_img_data()), places=4)
self.img_model.set_supersampling(2)
self.img_model.get_img_data()
def test_absorption_correction_with_different_image_sizes(self):
dummy_correction = DummyCorrection(self.img_model.get_img_data().shape, 0.4)
# self.img_data.set_absorption_correction(np.ones(self.img_data._img_data.shape)*0.4)
self.img_model.add_img_correction(dummy_correction, "Dummy 1")
self.assertTrue(self.img_model._img_corrections.has_items())
self.img_model.load(os.path.join(data_path, 'CeO2_Pilatus1M.tif'))
self.assertFalse(self.img_model.has_corrections())
def test_adding_several_absorption_corrections(self):
original_image = np.copy(self.img_model.get_img_data())
img_shape = original_image.shape
self.img_model.add_img_correction(DummyCorrection(img_shape, 0.4))
self.img_model.add_img_correction(DummyCorrection(img_shape, 3))
self.img_model.add_img_correction(DummyCorrection(img_shape, 5))
self.assertTrue(np.sum(original_image)/(0.5*3*5), np.sum(self.img_model.get_img_data()))
self.img_model.delete_img_correction(1)
self.assertTrue(np.sum(original_image)/(0.5*5), np.sum(self.img_model.get_img_data()))
def test_saving_data(self):
self.img_model.load(os.path.join(data_path, 'image_001.tif'))
filename = os.path.join(data_path, 'test.tif')
self.img_model.save(filename)
first_img_array = np.copy(self.img_model._img_data)
self.img_model.load(filename)
self.assertTrue(np.array_equal(first_img_array, self.img_model._img_data))
self.assertTrue(os.path.exists(filename))
os.remove(filename)
def test_rotation(self):
pre_transformed_data = self.img_model.get_img_data()
self.img_model.rotate_img_m90()
self.img_model.rotate_img_m90()
self.img_model.rotate_img_m90()
self.img_model.rotate_img_m90()
self.assertTrue(np.array_equal(self.img_model.get_img_data(), pre_transformed_data))
self.img_model.reset_img_transformations()
pre_transformed_data = self.img_model.get_img_data()
self.img_model.rotate_img_m90()
self.img_model.rotate_img_p90()
self.assertTrue(np.array_equal(self.img_model.get_img_data(), pre_transformed_data))
self.img_model.reset_img_transformations()
pre_transformed_data = self.img_model.get_img_data()
self.img_model.flip_img_horizontally()
self.img_model.flip_img_horizontally()
self.assertTrue(np.array_equal(self.img_model.get_img_data(), pre_transformed_data))
self.img_model.reset_img_transformations()
pre_transformed_data = self.img_model.get_img_data()
self.img_model.flip_img_vertically()
self.img_model.flip_img_vertically()
self.assertTrue(np.array_equal(self.img_model.get_img_data(), pre_transformed_data))
self.img_model.reset_img_transformations()
self.img_model.flip_img_vertically()
self.img_model.flip_img_horizontally()
self.img_model.rotate_img_m90()
self.img_model.rotate_img_p90()
self.img_model.rotate_img_m90()
self.img_model.rotate_img_m90()
self.img_model.flip_img_horizontally()
transformed_data = self.img_model.get_img_data()
self.img_model.load(os.path.join(data_path, 'image_001.tif'))
self.assertTrue(np.array_equal(self.img_model.get_img_data(), transformed_data))
self.img_model.reset_img_transformations()
pre_transformed_data = self.img_model.get_img_data()
self.img_model.rotate_img_m90()
self.img_model.reset_img_transformations()
self.assertTrue(np.array_equal(self.img_model.get_img_data(), pre_transformed_data))
pre_transformed_data = self.img_model.get_img_data()
self.img_model.rotate_img_p90()
self.img_model.reset_img_transformations()
self.assertTrue(np.array_equal(self.img_model.get_img_data(), pre_transformed_data))
pre_transformed_data = self.img_model.get_img_data()
self.img_model.flip_img_horizontally()
self.img_model.reset_img_transformations()
self.assertTrue(np.array_equal(self.img_model.get_img_data(), pre_transformed_data))
pre_transformed_data = self.img_model.get_img_data()
self.img_model.flip_img_vertically()
self.img_model.reset_img_transformations()
self.assertTrue(np.array_equal(self.img_model.get_img_data(), pre_transformed_data))
pre_transformed_data = self.img_model.get_img_data()
self.img_model.flip_img_vertically()
self.img_model.flip_img_horizontally()
self.img_model.rotate_img_m90()
self.img_model.rotate_img_p90()
self.img_model.rotate_img_m90()
self.img_model.rotate_img_m90()
self.img_model.flip_img_horizontally()
self.img_model.reset_img_transformations()
self.assertTrue(np.array_equal(self.img_model.get_img_data(), pre_transformed_data))
class ImgModelTest(unittest.TestCase):
@classmethod
def setUpClass(cls):
cls.app = QtGui.QApplication([])
@classmethod
def tearDownClass(cls):
cls.app.quit()
def setUp(self):
self.img_model = ImgModel()
self.img_model.load(os.path.join(data_path, 'image_001.tif'))
def tearDown(self):
del self.img_model
def perform_transformations_tests(self):
self.assertEqual(np.sum(np.absolute(self.img_model.get_img_data())), 0)
self.img_model.rotate_img_m90()
self.assertEqual(np.sum(np.absolute(self.img_model.get_img_data())), 0)
self.img_model.flip_img_horizontally()
self.assertEqual(np.sum(np.absolute(self.img_model.get_img_data())), 0)
self.img_model.rotate_img_p90()
self.assertEqual(np.sum(np.absolute(self.img_model.get_img_data())), 0)
self.img_model.flip_img_vertically()
self.assertEqual(np.sum(np.absolute(self.img_model.get_img_data())), 0)
self.img_model.reset_img_transformations()
self.assertEqual(np.sum(np.absolute(self.img_model.get_img_data())), 0)
def test_load_emits_signal(self):
callback_fcn = MagicMock()
self.img_model.img_changed.connect(callback_fcn)
self.img_model.load(os.path.join(data_path, 'image_001.tif'))
callback_fcn.assert_called_once_with()
def test_flipping_images(self):
original_image = np.copy(self.img_model._img_data)
self.img_model.flip_img_vertically()
self.assertTrue(np.array_equal(self.img_model._img_data, np.flipud(original_image)))
def test_simple_background_subtraction(self):
self.first_image = np.copy(self.img_model.get_img_data())
self.img_model.load_next_file()
self.second_image = np.copy(self.img_model.get_img_data())
self.img_model.load(os.path.join(data_path, 'image_001.tif'))
self.img_model.load_background(os.path.join(data_path, 'image_002.tif'))
self.assertFalse(np.array_equal(self.first_image, self.img_model.get_img_data()))
self.img_model.load_next_file()
self.assertEqual(np.sum(self.img_model.get_img_data()), 0)
def test_background_subtraction_with_supersampling(self):
self.img_model.load_background(os.path.join(data_path, 'image_002.tif'))
self.img_model.set_supersampling(2)
self.img_model.get_img_data()
self.img_model.set_supersampling(3)
self.img_model.get_img_data()
self.img_model.load_next_file()
self.img_model.get_img_data()
def test_background_subtraction_with_transformation(self):
self.img_model.load_background(os.path.join(data_path, 'image_002.tif'))
original_img = np.copy(self.img_model._img_data)
original_background = np.copy(self.img_model._background_data)
self.assertNotEqual(self.img_model._background_data, None)
self.assertFalse(np.array_equal(self.img_model.img_data, self.img_model._img_data))
original_img_background_subtracted = np.copy(self.img_model.get_img_data())
self.assertTrue(np.array_equal(original_img_background_subtracted, original_img - original_background))
### now comes the main process - flipping the image
self.img_model.flip_img_vertically()
flipped_img = np.copy(self.img_model._img_data)
self.assertTrue(np.array_equal(np.flipud(original_img), flipped_img))
flipped_background = np.copy(self.img_model._background_data)
self.assertTrue(np.array_equal(np.flipud(original_background), flipped_background))
flipped_img_background_subtracted = np.copy(self.img_model.get_img_data())
self.assertTrue(np.array_equal(flipped_img_background_subtracted, flipped_img - flipped_background))
self.assertTrue(np.array_equal(np.flipud(original_img_background_subtracted),
flipped_img_background_subtracted))
self.assertEqual(np.sum(np.flipud(original_img_background_subtracted) - flipped_img_background_subtracted), 0)
self.img_model.load(os.path.join(data_path, 'image_002.tif'))
self.perform_transformations_tests()
def test_background_subtraction_with_supersampling_and_image_transformation(self):
self.img_model.load_background(os.path.join(data_path, 'image_002.tif'))
self.img_model.load(os.path.join(data_path, 'image_002.tif'))
self.img_model.set_supersampling(2)
self.assertEqual(self.img_model.get_img_data().shape, (4096, 4096))
self.perform_transformations_tests()
self.img_model.set_supersampling(3)
self.assertEqual(self.img_model.get_img_data().shape, (6144, 6144))
self.perform_transformations_tests()
self.img_model.load(os.path.join(data_path, 'image_002.tif'))
self.assertEqual(self.img_model.get_img_data().shape, (6144, 6144))
self.perform_transformations_tests()
def test_background_scaling_and_offset(self):
self.img_model.load_background(os.path.join(data_path, 'image_002.tif'))
# assure that everything is correct before
self.assertTrue(np.array_equal(self.img_model.get_img_data(),
self.img_model._img_data - self.img_model._background_data))
# set scaling and see difference
self.img_model.set_background_scaling(2.4)
self.assertTrue(np.array_equal(self.img_model.get_img_data(),
self.img_model._img_data - 2.4 * self.img_model._background_data))
# set offset and see the difference
self.img_model.set_background_scaling(1.0)
self.img_model.set_background_offset(100.0)
self.assertTrue(np.array_equal(self.img_model.img_data,
self.img_model._img_data - (self.img_model._background_data + 100.0)))
# use offset and scaling combined
self.img_model.set_background_scaling(2.3)
self.img_model.set_background_offset(100.0)
self.assertTrue(np.array_equal(self.img_model.img_data,
self.img_model._img_data - (2.3 * self.img_model._background_data + 100)))
def test_background_with_different_shape(self):
self.img_model.load_background(os.path.join(data_path, 'CeO2_Pilatus1M.tif'))
self.assertEqual(self.img_model._background_data, None)
self.img_model.load_background(os.path.join(data_path, 'image_002.tif'))
self.assertTrue(self.img_model._background_data is not None)
self.img_model.load(os.path.join(data_path, 'CeO2_Pilatus1M.tif'))
self.assertEqual(self.img_model._background_data, None)
def test_absorption_correction_with_supersampling(self):
original_image = np.copy(self.img_model._img_data)
dummy_correction = DummyCorrection(self.img_model.get_img_data().shape, 0.6)
self.img_model.add_img_correction(dummy_correction, "Dummy 1")
self.assertAlmostEqual(np.sum(original_image / 0.6), np.sum(self.img_model.get_img_data()), places=4)
self.img_model.set_supersampling(2)
self.img_model.get_img_data()
def test_absorption_correction_with_different_image_sizes(self):
dummy_correction = DummyCorrection(self.img_model.get_img_data().shape, 0.4)
# self.img_data.set_absorption_correction(np.ones(self.img_data._img_data.shape)*0.4)
self.img_model.add_img_correction(dummy_correction, "Dummy 1")
self.assertTrue(self.img_model._img_corrections.has_items())
self.img_model.load(os.path.join(data_path, 'CeO2_Pilatus1M.tif'))
self.assertFalse(self.img_model.has_corrections())
def test_adding_several_absorption_corrections(self):
original_image = np.copy(self.img_model.get_img_data())
img_shape = original_image.shape
self.img_model.add_img_correction(DummyCorrection(img_shape, 0.4))
self.img_model.add_img_correction(DummyCorrection(img_shape, 3))
self.img_model.add_img_correction(DummyCorrection(img_shape, 5))
self.assertTrue(np.sum(original_image) / (0.5 * 3 * 5), np.sum(self.img_model.get_img_data()))
self.img_model.delete_img_correction(1)
self.assertTrue(np.sum(original_image) / (0.5 * 5), np.sum(self.img_model.get_img_data()))
def test_saving_data(self):
self.img_model.load(os.path.join(data_path, 'image_001.tif'))
filename = os.path.join(data_path, 'test.tif')
self.img_model.save(filename)
first_img_array = np.copy(self.img_model._img_data)
self.img_model.load(filename)
self.assertTrue(np.array_equal(first_img_array, self.img_model._img_data))
self.assertTrue(os.path.exists(filename))
os.remove(filename)
def test_negative_rotation(self):
pre_transformed_data = self.img_model.get_img_data()
self.img_model.rotate_img_m90()
self.img_model.rotate_img_m90()
self.img_model.rotate_img_m90()
self.img_model.rotate_img_m90()
self.assertTrue(np.array_equal(self.img_model.get_img_data(), pre_transformed_data))
def test_combined_rotation(self):
pre_transformed_data = self.img_model.get_img_data()
self.img_model.rotate_img_m90()
self.img_model.rotate_img_p90()
self.assertTrue(np.array_equal(self.img_model.get_img_data(), pre_transformed_data))
def test_flip_img_horizontally(self):
pre_transformed_data = self.img_model.get_img_data()
self.img_model.flip_img_horizontally()
self.img_model.flip_img_horizontally()
self.assertTrue(np.array_equal(self.img_model.get_img_data(), pre_transformed_data))
def test_flip_img_vertically(self):
pre_transformed_data = self.img_model.get_img_data()
self.img_model.flip_img_vertically()
self.img_model.flip_img_vertically()
self.assertTrue(np.array_equal(self.img_model.get_img_data(), pre_transformed_data))
def test_combined_rotation_and_flipping(self):
self.img_model.flip_img_vertically()
self.img_model.flip_img_horizontally()
self.img_model.rotate_img_m90()
self.img_model.rotate_img_p90()
self.img_model.rotate_img_m90()
self.img_model.rotate_img_m90()
self.img_model.flip_img_horizontally()
transformed_data = self.img_model.get_img_data()
self.img_model.load(os.path.join(data_path, 'image_001.tif'))
self.assertTrue(np.array_equal(self.img_model.get_img_data(), transformed_data))
def test_reset_img_transformation(self):
pre_transformed_data = self.img_model.get_img_data()
self.img_model.rotate_img_m90()
self.img_model.reset_img_transformations()
self.assertTrue(np.array_equal(self.img_model.get_img_data(), pre_transformed_data))
pre_transformed_data = self.img_model.get_img_data()
self.img_model.rotate_img_p90()
self.img_model.reset_img_transformations()
self.assertTrue(np.array_equal(self.img_model.get_img_data(), pre_transformed_data))
pre_transformed_data = self.img_model.get_img_data()
self.img_model.flip_img_horizontally()
self.img_model.reset_img_transformations()
self.assertTrue(np.array_equal(self.img_model.get_img_data(), pre_transformed_data))
pre_transformed_data = self.img_model.get_img_data()
self.img_model.flip_img_vertically()
self.img_model.reset_img_transformations()
self.assertTrue(np.array_equal(self.img_model.get_img_data(), pre_transformed_data))
pre_transformed_data = self.img_model.get_img_data()
self.img_model.flip_img_vertically()
self.img_model.flip_img_horizontally()
self.img_model.rotate_img_m90()
self.img_model.rotate_img_p90()
self.img_model.rotate_img_m90()
self.img_model.rotate_img_m90()
self.img_model.flip_img_horizontally()
self.img_model.reset_img_transformations()
self.assertTrue(np.array_equal(self.img_model.get_img_data(), pre_transformed_data))
def test_loading_a_tagged_tif_file_and_retrieving_info_string(self):
self.img_model.load(os.path.join(data_path, "attrib.tif"))
self.assertIn("areaDetector", self.img_model.file_info)
def test_loading_spe_file(self):
self.img_model.load(os.path.join(spe_path, 'CeO2_PI_CCD_Mo.SPE'))
self.assertEqual(self.img_model.img_data.shape, (1042, 1042))
