Replace all _FP/_FT direct calls.

Author: datumbox

torchvision/prototype/transforms/__init__.py

@@ -8,7 +8,7 @@
 from ._auto_augment import RandAugment, TrivialAugmentWide, AutoAugment
 from ._container import Compose, RandomApply, RandomChoice, RandomOrder
 from ._geometry import HorizontalFlip, Resize, CenterCrop, RandomResizedCrop
-from ._meta_conversion import ConvertBoundingBoxFormat, ConvertImageDtype, ConvertImageColorSpace
+from ._meta import ConvertBoundingBoxFormat, ConvertImageDtype, ConvertImageColorSpace
 from ._misc import Identity, Normalize, ToDtype, Lambda
 from ._presets import CocoEval, ImageNetEval, VocEval, Kinect400Eval, RaftEval
 from ._type_conversion import DecodeImage, LabelToOneHot

torchvision/prototype/transforms/_meta_conversion.py renamed to torchvision/prototype/transforms/_meta.py

@@ -4,7 +4,7 @@
 import torch
 from torchvision.prototype import features
 from torchvision.prototype.utils._internal import query_recursively
-from torchvision.transforms import functional_tensor as _FT, functional_pil as _FP
+from .functional._meta import get_dimensions_image_tensor, get_dimensions_image_pil
 
 
 def query_image(sample: Any) -> Union[PIL.Image.Image, torch.Tensor, features.Image]:
@@ -25,9 +25,9 @@ def get_image_dimensions(image: Union[PIL.Image.Image, torch.Tensor, features.Im
         channels = image.num_channels
         height, width = image.image_size
     elif isinstance(image, torch.Tensor):
-        channels, height, width = _FT.get_dimensions(image)
+        channels, height, width = get_dimensions_image_tensor(image)
     elif isinstance(image, PIL.Image.Image):
-        channels, height, width = _FP.get_dimensions(image)
+        channels, height, width = get_dimensions_image_pil(image)
     else:
         raise TypeError(f"unable to get image dimensions from object of type {type(image).__name__}")
     return channels, height, width

torchvision/prototype/transforms/_utils.py

@@ -1,5 +1,5 @@
 from torchvision.transforms import InterpolationMode  # usort: skip
-from ._meta_conversion import (
+from ._meta import (
     convert_bounding_box_format,
     convert_image_color_space_tensor,
     convert_image_color_space_pil,

torchvision/prototype/transforms/functional/__init__.py

@@ -8,7 +8,7 @@
 from torchvision.transforms import functional_tensor as _FT, functional_pil as _FP
 from torchvision.transforms.functional import pil_modes_mapping, _get_inverse_affine_matrix
 
-from ._meta_conversion import convert_bounding_box_format
+from ._meta import convert_bounding_box_format, get_dimensions_image_tensor, get_dimensions_image_pil
 
 
 horizontal_flip_image_tensor = _FT.hflip
@@ -39,7 +39,7 @@ def resize_image_tensor(
     antialias: Optional[bool] = None,
 ) -> torch.Tensor:
     new_height, new_width = size
-    num_channels, old_height, old_width = _FT.get_dimensions(image)
+    num_channels, old_height, old_width = get_dimensions_image_tensor(image)
     batch_shape = image.shape[:-3]
     return _FT.resize(
         image.reshape((-1, num_channels, old_height, old_width)),
@@ -141,7 +141,7 @@ def affine_image_tensor(
 
     center_f = [0.0, 0.0]
     if center is not None:
-        _, height, width = _FT.get_dimensions(img)
+        _, height, width = get_dimensions_image_tensor(img)
         # Center values should be in pixel coordinates but translated such that (0, 0) corresponds to image center.
         center_f = [1.0 * (c - s * 0.5) for c, s in zip(center, (width, height))]
 
@@ -167,7 +167,7 @@ def affine_image_pil(
     # it is visually better to estimate the center without 0.5 offset
     # otherwise image rotated by 90 degrees is shifted vs output image of torch.rot90 or F_t.affine
     if center is None:
-        _, height, width = _FP.get_dimensions(img)
+        _, height, width = get_dimensions_image_pil(img)
         center = [width * 0.5, height * 0.5]
     matrix = _get_inverse_affine_matrix(center, angle, translate, scale, shear)
 
@@ -184,7 +184,7 @@ def rotate_image_tensor(
 ) -> torch.Tensor:
     center_f = [0.0, 0.0]
     if center is not None:
-        _, height, width = _FT.get_dimensions(img)
+        _, height, width = get_dimensions_image_tensor(img)
         # Center values should be in pixel coordinates but translated such that (0, 0) corresponds to image center.
         center_f = [1.0 * (c - s * 0.5) for c, s in zip(center, (width, height))]
 
@@ -260,13 +260,13 @@ def _center_crop_compute_crop_anchor(
 
 def center_crop_image_tensor(img: torch.Tensor, output_size: List[int]) -> torch.Tensor:
     crop_height, crop_width = _center_crop_parse_output_size(output_size)
-    _, image_height, image_width = _FT.get_dimensions(img)
+    _, image_height, image_width = get_dimensions_image_tensor(img)
 
     if crop_height > image_height or crop_width > image_width:
         padding_ltrb = _center_crop_compute_padding(crop_height, crop_width, image_height, image_width)
         img = pad_image_tensor(img, padding_ltrb, fill=0)
 
-        _, image_height, image_width = _FT.get_dimensions(img)
+        _, image_height, image_width = get_dimensions_image_tensor(img)
         if crop_width == image_width and crop_height == image_height:
             return img
 
@@ -276,13 +276,13 @@ def center_crop_image_tensor(img: torch.Tensor, output_size: List[int]) -> torch
 
 def center_crop_image_pil(img: PIL.Image.Image, output_size: List[int]) -> PIL.Image.Image:
     crop_height, crop_width = _center_crop_parse_output_size(output_size)
-    _, image_height, image_width = _FP.get_dimensions(img)
+    _, image_height, image_width = get_dimensions_image_pil(img)
 
     if crop_height > image_height or crop_width > image_width:
         padding_ltrb = _center_crop_compute_padding(crop_height, crop_width, image_height, image_width)
         img = pad_image_pil(img, padding_ltrb, fill=0)
 
-        _, image_height, image_width = _FP.get_dimensions(img)
+        _, image_height, image_width = get_dimensions_image_pil(img)
         if crop_width == image_width and crop_height == image_height:
             return img
 

torchvision/prototype/transforms/functional/_geometry.py

@@ -4,6 +4,10 @@
 from torchvision.transforms import functional_tensor as _FT, functional_pil as _FP
 
 
+get_dimensions_image_tensor = _FT.get_dimensions
+get_dimensions_image_pil = _FP.get_dimensions
+
+
 def _xywh_to_xyxy(xywh: torch.Tensor) -> torch.Tensor:
     xyxy = xywh.clone()
     xyxy[..., 2:] += xyxy[..., :2]

torchvision/prototype/transforms/functional/_meta_conversion.py renamed to torchvision/prototype/transforms/functional/_meta.py

@@ -2,13 +2,10 @@
 
 import PIL.Image
 import torch
-from torchvision.transforms import functional_tensor as _FT, functional_pil as _FP
+from torchvision.transforms import functional_tensor as _FT
 from torchvision.transforms.functional import to_tensor, to_pil_image
 
 
-get_dimensions_image_tensor = _FT.get_dimensions
-get_dimensions_image_pil = _FP.get_dimensions
-
 normalize_image_tensor = _FT.normalize