fixes docstrings

Signed-off-by: Wenqi Li <wenqil@nvidia.com>

Author: wyli

monai/transforms/compose.py

@@ -43,7 +43,7 @@ def eval_lazy_stack(
     data, upcoming, lazy_evaluation: bool = False, mode=GridSampleMode.BILINEAR, padding_mode=GridSamplePadMode.BORDER
 ):
     """
-    Given the upcoming transform ``upcoming``, if lazy_resample is True, go through the Metatensors and
+    Given the upcoming transform ``upcoming``, if lazy_resample is True, go through the MetaTensors and
     evaluate the lazy applied operations. The returned `data` will then be ready for the ``upcoming`` transform.
     """
     if not lazy_evaluation:

monai/transforms/inverse.py

@@ -156,6 +156,7 @@ def push_pending_transform(
     ) -> None:
         """
         Push to MetaTensor's pending operations for later execution.
+
         Args:
             data:
             key:

monai/transforms/spatial/array.py

@@ -839,7 +839,7 @@ def lazy_call(self, img, axes) -> torch.Tensor:
         spatial_chn_shape = [1, *convert_to_numpy(_shape, wrap_sequence=True).tolist()]
         _affine = self.update_meta(img, spatial_chn_shape, axes)
         self.push_pending_transform(img, lazy_shape=_shape, lazy_affine=_affine)
-        return img  # type: ignore
+        return img
 
     def __call__(self, img: torch.Tensor) -> torch.Tensor:
         """
@@ -970,7 +970,7 @@ def __call__(
 
         _mode = look_up_option(self.mode if mode is None else mode, InterpolateMode)
         _align_corners = self.align_corners if align_corners is None else align_corners
-        img = convert_to_tensor(img, track_meta=get_track_meta())  # type: ignore
+        img = convert_to_tensor(img, track_meta=get_track_meta())
         original_sp_size = img.peek_pending_shape() if isinstance(img, MetaTensor) else img.shape[1:]
         if self.lazy_evaluation:
             if anti_aliasing:
@@ -1192,7 +1192,7 @@ def lazy_call(self, img, output_shape, transform_t, mode, padding_mode, align_co
                 "dtype": str(dtype)[6:],
             },
         )
-        return img  # type: ignore
+        return img
 
     def update_meta(self, img, rotate_mat):
         affine = convert_to_tensor(img.peek_pending_affine(), track_meta=False)
@@ -1370,7 +1370,7 @@ def lazy_call(self, img, zoom_size, mode, align_corners) -> torch.Tensor:
                 "padcrop": {},
             },
         )
-        return img  # type: ignore
+        return img
 
     def inverse(self, data: torch.Tensor) -> torch.Tensor:
         transform = self.pop_transform(data)

monai/transforms/spatial/dictionary.py

@@ -920,7 +920,7 @@ def __call__(self, data: Mapping[Hashable, NdarrayOrTensor]) -> dict[Hashable, N
         self.rand_affine.randomize()
 
         item = d[first_key]
-        spatial_size = item.peek_pending_shape() if isinstance(item, MetaTensor) else item.shape[1:]  # type: ignore
+        spatial_size = item.peek_pending_shape() if isinstance(item, MetaTensor) else item.shape[1:]
 
         sp_size = fall_back_tuple(self.rand_affine.spatial_size, spatial_size)
         # change image size or do random transform