Fix, better tests, expose more stuff

Author: NicolasHug

test/test_transforms_v2.py

@@ -6171,50 +6171,48 @@ def test_transform_invalid_quality_error(self, quality):
             transforms.JPEG(quality=quality)
 
 
-class TestQuerySize:
+class TestUtils:
+    # TODO: Still need to test has_all, has_any, check_type and get_bouding_boxes
     @pytest.mark.parametrize(
-        "make_input, input_name",
-        [
-            (lambda: torch.rand(3, 32, 64), "pure_tensor"),
-            (lambda: tv_tensors.Image(torch.rand(3, 32, 64)), "tv_tensor_image"),
-            (lambda: PIL.Image.new("RGB", (64, 32)), "pil_image"),
-            (lambda: tv_tensors.Video(torch.rand(1, 3, 32, 64)), "tv_tensor_video"),
-            (lambda: tv_tensors.Mask(torch.randint(0, 2, (32, 64))), "tv_tensor_mask"),
-        ],
-        ids=["pure_tensor", "tv_tensor_image", "pil_image", "tv_tensor_video", "tv_tensor_mask"],
+        "make_input1", [make_image_tensor, make_image_pil, make_image, make_bounding_boxes, make_segmentation_mask]
+    )
+    @pytest.mark.parametrize(
+        "make_input2", [make_image_tensor, make_image_pil, make_image, make_bounding_boxes, make_segmentation_mask]
     )
-    def test_functional(self, make_input, input_name):
-        input1 = make_input()
-        input2 = make_input()
-        # Both inputs should have the same size (32, 64)
-        assert transforms.query_size([input1, input2]) == (32, 64)
+    @pytest.mark.parametrize("query", [transforms.query_size, transforms.query_chw])
+    def test_query_size_and_query_chw(self, make_input1, make_input2, query):
+        size = (32, 64)
+        input1 = make_input1(size)
+        input2 = make_input2(size)
+
+        if query is transforms.query_chw and not any(
+            transforms.check_type(inpt, (is_pure_tensor, tv_tensors.Image, PIL.Image.Image, tv_tensors.Video))
+            for inpt in (input1, input2)
+        ):
+            return
+
+        expected = size if query is transforms.query_size else ((3,) + size)
+        assert query([input1, input2]) == expected
 
     @pytest.mark.parametrize(
-        "make_input, input_name",
-        [
-            (lambda: torch.rand(3, 32, 64), "pure_tensor"),
-            (lambda: tv_tensors.Image(torch.rand(3, 32, 64)), "tv_tensor_image"),
-            (lambda: PIL.Image.new("RGB", (64, 32)), "pil_image"),
-            (lambda: tv_tensors.Video(torch.rand(1, 3, 32, 64)), "tv_tensor_video"),
-            (lambda: tv_tensors.Mask(torch.randint(0, 2, (32, 64))), "tv_tensor_mask"),
-        ],
-        ids=["pure_tensor", "tv_tensor_image", "pil_image", "tv_tensor_video", "tv_tensor_mask"],
-    )
-    def test_functional_mixed_types(self, make_input, input_name):
-        input1 = make_input()
-        input2 = make_input()
-        # Both inputs should have the same size (32, 64)
-        assert transforms.query_size([input1, input2]) == (32, 64)
-
-    def test_different_sizes(self):
-        img_tensor = torch.rand(3, 32, 64)  # (C, H, W)
-        img_tensor_different_size = torch.rand(3, 48, 96)  # (C, H, W)
-        # Should raise ValueError for different sizes
-        with pytest.raises(ValueError, match="Found multiple HxW dimensions"):
-            transforms.query_size([img_tensor, img_tensor_different_size])
-
-    def test_no_valid_image(self):
-        invalid_input = torch.rand(1, 10)  # Non-image tensor
-        # Should raise TypeError for invalid input
-        with pytest.raises(TypeError, match="No image, video, mask or bounding box was found"):
-            transforms.query_size([invalid_input])
+        "make_input1", [make_image_tensor, make_image_pil, make_image, make_bounding_boxes, make_segmentation_mask]
+    )
+    @pytest.mark.parametrize(
+        "make_input2", [make_image_tensor, make_image_pil, make_image, make_bounding_boxes, make_segmentation_mask]
+    )
+    @pytest.mark.parametrize("query", [transforms.query_size, transforms.query_chw])
+    def test_different_sizes(self, make_input1, make_input2, query):
+        input1 = make_input1((10, 10))
+        input2 = make_input2((20, 20))
+        if query is transforms.query_chw and not all(
+            transforms.check_type(inpt, (is_pure_tensor, tv_tensors.Image, PIL.Image.Image, tv_tensors.Video))
+            for inpt in (input1, input2)
+        ):
+            return
+        with pytest.raises(ValueError, match="Found multiple"):
+            query([input1, input2])
+
+    @pytest.mark.parametrize("query", [transforms.query_size, transforms.query_chw])
+    def test_no_valid_input(self, query):
+        with pytest.raises(TypeError, match="No image"):
+            query(["blah"])

torchvision/transforms/v2/__init__.py

@@ -55,5 +55,6 @@
 )
 from ._temporal import UniformTemporalSubsample
 from ._type_conversion import PILToTensor, ToImage, ToPILImage, ToPureTensor
+from ._utils import check_type, get_bounding_boxes, has_all, has_any, query_chw, query_size
 
 from ._deprecated import ToTensor  # usort: skip

torchvision/transforms/v2/functional/__init__.py

@@ -151,5 +151,3 @@
 from ._type_conversion import pil_to_tensor, to_image, to_pil_image
 
 from ._deprecated import get_image_size, to_tensor  # usort: skip
-
-from ._utils import query_size