Remove old asserts for python 3.12 support

keras_cv/src/datasets/pascal_voc/segmentation_test.py

@@ -105,13 +105,13 @@ def test_get_image_ids(self):
         train_ids = ["2007_000032", "2007_000039", "2007_000063"]
         eval_ids = ["2007_000033"]
         train_eval_ids = train_ids + eval_ids
-        self.assertEquals(
+        self.assertEqual(
             segmentation._get_image_ids(data_dir, "train"), train_ids
         )
-        self.assertEquals(
+        self.assertEqual(
             segmentation._get_image_ids(data_dir, "eval"), eval_ids
         )
-        self.assertEquals(
+        self.assertEqual(
             segmentation._get_image_ids(data_dir, "trainval"), train_eval_ids
         )
 
@@ -161,7 +161,7 @@ def test_parse_annotation_file(self):
                 },
             ],
         }
-        self.assertEquals(metadata, expected_result)
+        self.assertEqual(metadata, expected_result)
 
     def test_decode_png_mask(self):
         local_data_dir = os.path.join(self.tempdir, "pascal_voc_2012/")
@@ -177,19 +177,19 @@ def test_decode_png_mask(self):
         segmentation._maybe_populate_voc_color_mapping()
         mask = segmentation._decode_png_mask(mask)
 
-        self.assertEquals(mask.shape, (281, 500, 1))
-        self.assertEquals(
+        self.assertEqual(mask.shape, (281, 500, 1))
+        self.assertEqual(
             tf.reduce_max(mask), 255
         )  # The 255 value is for the boundary
-        self.assertEquals(
+        self.assertEqual(
             tf.reduce_min(mask), 0
         )  # The 0 value is for the background
         # The mask contains two classes, 1 and 15, see the label section in the
         # previous test case.
-        self.assertEquals(
+        self.assertEqual(
             tf.reduce_sum(tf.cast(tf.equal(mask, 1), tf.int32)), 4734
         )
-        self.assertEquals(
+        self.assertEqual(
             tf.reduce_sum(tf.cast(tf.equal(mask, 15), tf.int32)), 866
         )
 
@@ -245,7 +245,7 @@ def test_parse_single_image(self):
                 data_dir, "SegmentationObject", "2007_000032.png"
             ),
         }
-        self.assertEquals(result_dict, expected_result)
+        self.assertEqual(result_dict, expected_result)
 
     def test_build_metadata(self):
         local_data_dir = os.path.join(self.tempdir, "pascal_voc_2012/")
@@ -257,7 +257,7 @@ def test_build_metadata(self):
         image_ids = segmentation._get_image_ids(data_dir, "trainval")
         metadata = segmentation._build_metadata(data_dir, image_ids)
 
-        self.assertEquals(
+        self.assertEqual(
             metadata["image/filename"],
             [
                 "2007_000032.jpg",
@@ -296,7 +296,7 @@ def test_build_dataset(self):
         dataset = segmentation._build_dataset_from_metadata(metadata)
 
         entry = next(dataset.take(1).as_numpy_iterator())
-        self.assertEquals(entry["image/filename"], b"2007_000032.jpg")
+        self.assertEqual(entry["image/filename"], b"2007_000032.jpg")
         expected_keys = [
             "image",
             "image/filename",
@@ -316,18 +316,18 @@ def test_build_dataset(self):
 
         # Check the mask png content
         png = entry["class_segmentation"]
-        self.assertEquals(png.shape, (281, 500, 1))
-        self.assertEquals(
+        self.assertEqual(png.shape, (281, 500, 1))
+        self.assertEqual(
             tf.reduce_max(png), 255
         )  # The 255 value is for the boundary
-        self.assertEquals(
+        self.assertEqual(
             tf.reduce_min(png), 0
         )  # The 0 value is for the background
         # The mask contains two classes, 1 and 15, see the label section in the
         # previous test case.
-        self.assertEquals(
+        self.assertEqual(
             tf.reduce_sum(tf.cast(tf.equal(png, 1), tf.int32)), 4734
         )
-        self.assertEquals(
+        self.assertEqual(
             tf.reduce_sum(tf.cast(tf.equal(png, 15), tf.int32)), 866
         )

keras_cv/src/datasets/waymo/load_test.py

@@ -44,7 +44,7 @@ def test_load_from_directory(self):
         # Extract records into a list
         dataset = [record for record in dataset]
 
-        self.assertEquals(len(dataset), 1)
+        self.assertEqual(len(dataset), 1)
         self.assertNotEqual(dataset[0]["timestamp_micros"], 0)
 
     @pytest.mark.skipif(
@@ -58,5 +58,5 @@ def test_load_from_files(self):
         # Extract records into a list
         dataset = [record for record in dataset]
 
-        self.assertEquals(len(dataset), 1)
+        self.assertEqual(len(dataset), 1)
         self.assertNotEqual(dataset[0]["timestamp_micros"], 0)

keras_cv/src/layers/augmenter_test.py

@@ -27,7 +27,7 @@ def test_call(self):
             ]
         )
         output = augmenter(images)
-        self.assertEquals(output.shape, images.shape)
+        self.assertEqual(output.shape, images.shape)
 
     def test_call_with_labels(self):
         images = {
@@ -43,4 +43,4 @@ def test_call_with_labels(self):
             ]
         )
         output = augmenter(images)
-        self.assertEquals(output["images"].shape, images["images"].shape)
+        self.assertEqual(output["images"].shape, images["images"].shape)

keras_cv/src/layers/feature_pyramid_test.py

@@ -30,7 +30,7 @@ def test_return_type_dict(self):
         inputs = {2: c2, 3: c3, 4: c4, 5: c5}
         output = layer(inputs)
         self.assertTrue(isinstance(output, dict))
-        self.assertEquals(sorted(output.keys()), [2, 3, 4, 5])
+        self.assertEqual(sorted(output.keys()), [2, 3, 4, 5])
 
     def test_result_shapes(self):
         layer = FeaturePyramid(min_level=2, max_level=5)
@@ -42,9 +42,9 @@ def test_result_shapes(self):
         inputs = {2: c2, 3: c3, 4: c4, 5: c5}
         output = layer(inputs)
         for level in inputs.keys():
-            self.assertEquals(output[level].shape[1], inputs[level].shape[1])
-            self.assertEquals(output[level].shape[2], inputs[level].shape[2])
-            self.assertEquals(output[level].shape[3], layer.num_channels)
+            self.assertEqual(output[level].shape[1], inputs[level].shape[1])
+            self.assertEqual(output[level].shape[2], inputs[level].shape[2])
+            self.assertEqual(output[level].shape[3], layer.num_channels)
 
         # Test with different resolution and channel size
         c2 = np.ones([2, 64, 128, 4])
@@ -56,9 +56,9 @@ def test_result_shapes(self):
         layer = FeaturePyramid(min_level=2, max_level=5)
         output = layer(inputs)
         for level in inputs.keys():
-            self.assertEquals(output[level].shape[1], inputs[level].shape[1])
-            self.assertEquals(output[level].shape[2], inputs[level].shape[2])
-            self.assertEquals(output[level].shape[3], layer.num_channels)
+            self.assertEqual(output[level].shape[1], inputs[level].shape[1])
+            self.assertEqual(output[level].shape[2], inputs[level].shape[2])
+            self.assertEqual(output[level].shape[3], layer.num_channels)
 
     def test_with_keras_input_tensor(self):
         # This mimic the model building with Backbone network
@@ -71,13 +71,13 @@ def test_with_keras_input_tensor(self):
         inputs = {2: c2, 3: c3, 4: c4, 5: c5}
         output = layer(inputs)
         for level in inputs.keys():
-            self.assertEquals(output[level].shape[1], inputs[level].shape[1])
-            self.assertEquals(output[level].shape[2], inputs[level].shape[2])
-            self.assertEquals(output[level].shape[3], layer.num_channels)
+            self.assertEqual(output[level].shape[1], inputs[level].shape[1])
+            self.assertEqual(output[level].shape[2], inputs[level].shape[2])
+            self.assertEqual(output[level].shape[3], layer.num_channels)
 
     def test_invalid_lateral_layers(self):
         lateral_layers = [keras.layers.Conv2D(256, 1)] * 3
-        with self.assertRaisesRegexp(
+        with self.assertRaisesRegex(
             ValueError, "Expect lateral_layers to be a dict"
         ):
             _ = FeaturePyramid(
@@ -88,7 +88,7 @@ def test_invalid_lateral_layers(self):
             3: keras.layers.Conv2D(256, 1),
             4: keras.layers.Conv2D(256, 1),
         }
-        with self.assertRaisesRegexp(
+        with self.assertRaisesRegex(
             ValueError, "with keys as .* [2, 3, 4, 5]"
         ):
             _ = FeaturePyramid(
@@ -97,7 +97,7 @@ def test_invalid_lateral_layers(self):
 
     def test_invalid_output_layers(self):
         output_layers = [keras.layers.Conv2D(256, 3)] * 3
-        with self.assertRaisesRegexp(
+        with self.assertRaisesRegex(
             ValueError, "Expect output_layers to be a dict"
         ):
             _ = FeaturePyramid(
@@ -108,7 +108,7 @@ def test_invalid_output_layers(self):
             3: keras.layers.Conv2D(256, 3),
             4: keras.layers.Conv2D(256, 3),
         }
-        with self.assertRaisesRegexp(
+        with self.assertRaisesRegex(
             ValueError, "with keys as .* [2, 3, 4, 5]"
         ):
             _ = FeaturePyramid(
@@ -126,13 +126,13 @@ def test_invalid_input_features(self):
         # Build required for Keas 3
         _ = layer(inputs)
         list_input = [c2, c3, c4, c5]
-        with self.assertRaisesRegexp(
+        with self.assertRaisesRegex(
             ValueError, "expects input features to be a dict"
         ):
             layer(list_input)
 
         dict_input_with_missing_feature = {2: c2, 3: c3, 4: c4}
-        with self.assertRaisesRegexp(
+        with self.assertRaisesRegex(
             ValueError, "Expect feature keys.*[2, 3, 4, 5]"
         ):
             layer(dict_input_with_missing_feature)

keras_cv/src/layers/fusedmbconv_test.py

@@ -36,7 +36,7 @@ def test_same_input_output_shapes(self):
         layer = FusedMBConvBlock(input_filters=32, output_filters=32)
 
         output = layer(inputs)
-        self.assertEquals(output.shape, [1, 64, 64, 32])
+        self.assertEqual(output.shape, [1, 64, 64, 32])
         self.assertLen(output, 1)
         self.assertTrue(isinstance(output, tf.Tensor))
 
@@ -45,7 +45,7 @@ def test_different_input_output_shapes(self):
         layer = FusedMBConvBlock(input_filters=32, output_filters=48)
 
         output = layer(inputs)
-        self.assertEquals(output.shape, [1, 64, 64, 48])
+        self.assertEqual(output.shape, [1, 64, 64, 48])
         self.assertLen(output, 1)
         self.assertTrue(isinstance(output, tf.Tensor))
 
@@ -56,6 +56,6 @@ def test_squeeze_excitation_ratio(self):
         )
 
         output = layer(inputs)
-        self.assertEquals(output.shape, [1, 64, 64, 48])
+        self.assertEqual(output.shape, [1, 64, 64, 48])
         self.assertLen(output, 1)
         self.assertTrue(isinstance(output, tf.Tensor))

keras_cv/src/layers/mbconv_test.py

@@ -36,7 +36,7 @@ def test_same_input_output_shapes(self):
         layer = MBConvBlock(input_filters=32, output_filters=32)
 
         output = layer(inputs)
-        self.assertEquals(output.shape, [1, 64, 64, 32])
+        self.assertEqual(output.shape, [1, 64, 64, 32])
         self.assertLen(output, 1)
         self.assertTrue(isinstance(output, tf.Tensor))
 
@@ -45,7 +45,7 @@ def test_different_input_output_shapes(self):
         layer = MBConvBlock(input_filters=32, output_filters=48)
 
         output = layer(inputs)
-        self.assertEquals(output.shape, [1, 64, 64, 48])
+        self.assertEqual(output.shape, [1, 64, 64, 48])
         self.assertLen(output, 1)
         self.assertTrue(isinstance(output, tf.Tensor))
 
@@ -54,6 +54,6 @@ def test_squeeze_excitation_ratio(self):
         layer = MBConvBlock(input_filters=32, output_filters=48, se_ratio=0.25)
 
         output = layer(inputs)
-        self.assertEquals(output.shape, [1, 64, 64, 48])
+        self.assertEqual(output.shape, [1, 64, 64, 48])
         self.assertLen(output, 1)
         self.assertTrue(isinstance(output, tf.Tensor))

keras_cv/src/layers/preprocessing/cut_mix_test.py

@@ -236,7 +236,7 @@ def test_single_image_input(self):
         ys = tf.one_hot(tf.constant([1]), 2)
         inputs = {"images": xs, "labels": ys}
         layer = CutMix()
-        with self.assertRaisesRegexp(
+        with self.assertRaisesRegex(
             ValueError, "CutMix received a single image to `call`"
         ):
             _ = layer(inputs)
@@ -246,15 +246,15 @@ def test_int_labels(self):
         ys = tf.one_hot(tf.constant([1, 0]), 2, dtype=tf.int32)
         inputs = {"images": xs, "labels": ys}
         layer = CutMix()
-        with self.assertRaisesRegexp(
+        with self.assertRaisesRegex(
             ValueError, "CutMix received labels with type"
         ):
             _ = layer(inputs)
 
     def test_image_input(self):
         xs = tf.ones((2, 512, 512, 3))
         layer = CutMix()
-        with self.assertRaisesRegexp(
+        with self.assertRaisesRegex(
             ValueError, "CutMix expects inputs in a dictionary with format"
         ):
             _ = layer(xs)

keras_cv/src/layers/preprocessing/fourier_mix_test.py

@@ -147,7 +147,7 @@ def test_image_input_only(self):
             tf.float32,
         )
         layer = FourierMix()
-        with self.assertRaisesRegexp(
+        with self.assertRaisesRegex(
             ValueError, "expects inputs in a dictionary"
         ):
             _ = layer(xs)
@@ -157,7 +157,7 @@ def test_single_image_input(self):
         ys = tf.one_hot(tf.constant([1]), 2)
         inputs = {"images": xs, "labels": ys}
         layer = FourierMix()
-        with self.assertRaisesRegexp(
+        with self.assertRaisesRegex(
             ValueError, "FourierMix received a single image to `call`"
         ):
             _ = layer(inputs)

keras_cv/src/layers/preprocessing/mix_up_test.py

@@ -159,7 +159,7 @@ def test_image_input_only(self):
             tf.float32,
         )
         layer = MixUp()
-        with self.assertRaisesRegexp(
+        with self.assertRaisesRegex(
             ValueError, "expects inputs in a dictionary"
         ):
             _ = layer(xs)
@@ -169,7 +169,7 @@ def test_single_image_input(self):
         ys = tf.one_hot(tf.constant([1]), 2)
         inputs = {"images": xs, "labels": ys}
         layer = MixUp()
-        with self.assertRaisesRegexp(
+        with self.assertRaisesRegex(
             ValueError, "MixUp received a single image to `call`"
         ):
             _ = layer(inputs)
@@ -179,15 +179,15 @@ def test_int_labels(self):
         ys = tf.one_hot(tf.constant([1, 0]), 2, dtype=tf.int32)
         inputs = {"images": xs, "labels": ys}
         layer = MixUp()
-        with self.assertRaisesRegexp(
+        with self.assertRaisesRegex(
             ValueError, "MixUp received labels with type"
         ):
             _ = layer(inputs)
 
     def test_image_input(self):
         xs = tf.ones((2, 512, 512, 3))
         layer = MixUp()
-        with self.assertRaisesRegexp(
+        with self.assertRaisesRegex(
             ValueError, "MixUp expects inputs in a dictionary with format"
         ):
             _ = layer(xs)

keras_cv/src/layers/preprocessing/mosaic_test.py

@@ -91,7 +91,7 @@ def test_image_input_only(self):
             tf.float32,
         )
         layer = Mosaic()
-        with self.assertRaisesRegexp(
+        with self.assertRaisesRegex(
             ValueError, "expects inputs in a dictionary"
         ):
             _ = layer(xs)
@@ -101,15 +101,15 @@ def test_single_image_input(self):
         ys = tf.one_hot(tf.constant([1]), 2)
         inputs = {"images": xs, "labels": ys}
         layer = Mosaic()
-        with self.assertRaisesRegexp(
+        with self.assertRaisesRegex(
             ValueError, "Mosaic received a single image to `call`"
         ):
             _ = layer(inputs)
 
     def test_image_input(self):
         xs = tf.ones((2, 512, 512, 3))
         layer = Mosaic()
-        with self.assertRaisesRegexp(
+        with self.assertRaisesRegex(
             ValueError, "Mosaic expects inputs in a dictionary with format"
         ):
             _ = layer(xs)