2418 Fix mypy errors caused by numpy(1.21.0) (#2419)

* [DLMED] test mypy fix

Signed-off-by: Nic Ma <[email protected]>

* [DLMED] test commit

Signed-off-by: Nic Ma <[email protected]>

* [DLMED] fix np.random

Signed-off-by: Nic Ma <[email protected]>

* [DLMED] fix other numpy data type errors

Signed-off-by: Nic Ma <[email protected]>

Author: Nic-Ma

monai/data/synthetic.py

@@ -48,7 +48,7 @@ def create_test_image_2d(
         random_state: the random generator to use. Defaults to `np.random`.
     """
     image = np.zeros((width, height))
-    rs = np.random if random_state is None else random_state
+    rs: np.random.RandomState = np.random.random.__self__ if random_state is None else random_state  # type: ignore
 
     for _ in range(num_objs):
         x = rs.randint(rad_max, width - rad_max)
@@ -111,7 +111,7 @@ def create_test_image_3d(
         :py:meth:`~create_test_image_2d`
     """
     image = np.zeros((width, height, depth))
-    rs = np.random if random_state is None else random_state
+    rs: np.random.RandomState = np.random.random.__self__ if random_state is None else random_state  # type: ignore
 
     for _ in range(num_objs):
         x = rs.randint(rad_max, width - rad_max)

monai/data/utils.py

@@ -605,7 +605,7 @@ def to_affine_nd(r: Union[np.ndarray, int], affine: np.ndarray) -> np.ndarray:
         raise ValueError(f"affine must have 2 dimensions, got {affine_np.ndim}.")
     new_affine = np.array(r, dtype=np.float64, copy=True)
     if new_affine.ndim == 0:
-        sr = new_affine.astype(int)
+        sr: int = int(new_affine.astype(np.uint))
         if not np.isfinite(sr) or sr < 0:
             raise ValueError(f"r must be positive, got {sr}.")
         new_affine = np.eye(sr + 1, dtype=np.float64)

monai/handlers/utils.py

@@ -206,12 +206,12 @@ class    mean    median    max    5percentile 95percentile  notnans
             if summary_ops is not None:
                 supported_ops = OrderedDict(
                     {
-                        "mean": np.nanmean,
-                        "median": np.nanmedian,
-                        "max": np.nanmax,
-                        "min": np.nanmin,
+                        "mean": lambda x: np.nanmean(x),
+                        "median": lambda x: np.nanmedian(x),
+                        "max": lambda x: np.nanmax(x),
+                        "min": lambda x: np.nanmin(x),
                         "90percentile": lambda x: np.nanpercentile(x[0], x[1]),
-                        "std": np.nanstd,
+                        "std": lambda x: np.nanstd(x),
                         "notnans": lambda x: (~np.isnan(x)).sum(),
                     }
                 )

monai/transforms/intensity/array.py

@@ -75,7 +75,7 @@ def __init__(self, prob: float = 0.1, mean: Union[Sequence[float], float] = 0.0,
         RandomizableTransform.__init__(self, prob)
         self.mean = mean
         self.std = std
-        self._noise = None
+        self._noise: np.ndarray
 
     def randomize(self, im_shape: Sequence[int]) -> None:
         super().randomize(None)
@@ -133,8 +133,8 @@ def __init__(
         self.channel_wise = channel_wise
         self.relative = relative
         self.sample_std = sample_std
-        self._noise1 = None
-        self._noise2 = None
+        self._noise1: np.ndarray
+        self._noise2: np.ndarray
 
     def _add_noise(self, img: Union[torch.Tensor, np.ndarray], mean: float, std: float):
         im_shape = img.shape
@@ -463,7 +463,7 @@ def randomize(self, data: np.ndarray) -> None:
         self.spatial_shape = data.shape[1:]
         self.rank = len(self.spatial_shape)
         n_coeff = int(np.prod([(self.degree + k) / k for k in range(1, self.rank + 1)]))
-        self._coeff = self.R.uniform(*self.coeff_range, n_coeff)
+        self._coeff = self.R.uniform(*self.coeff_range, n_coeff).tolist()
 
     def __call__(self, img: np.ndarray):
         """
@@ -670,13 +670,13 @@ def __init__(self, prob: float = 0.1, gamma: Union[Sequence[float], float] = (0.
                 raise AssertionError("gamma should be a number or pair of numbers.")
             self.gamma = (min(gamma), max(gamma))
 
-        self.gamma_value = None
+        self.gamma_value: float
 
     def randomize(self, data: Optional[Any] = None) -> None:
         super().randomize(None)
         self.gamma_value = self.R.uniform(low=self.gamma[0], high=self.gamma[1])
 
-    def __call__(self, img: np.ndarray) -> np.ndarray:
+    def __call__(self, img: np.ndarray):
         """
         Apply the transform to `img`.
         """

monai/transforms/spatial/array.py

@@ -464,7 +464,7 @@ def __call__(
             raise ValueError(f"Unsupported img dimension: {input_ndim}, available options are [2, 3].")
         _angle = ensure_tuple_rep(self.angle, 1 if input_ndim == 2 else 3)
         transform = create_rotate(input_ndim, _angle)
-        shift = create_translate(input_ndim, (im_shape - 1) / 2)
+        shift = create_translate(input_ndim, ((im_shape - 1) / 2).tolist())
         if self.keep_size:
             output_shape = im_shape
         else:
@@ -473,7 +473,7 @@ def __call__(
             )
             corners = transform[:-1, :-1] @ corners
             output_shape = np.asarray(corners.ptp(axis=1) + 0.5, dtype=int)
-        shift_1 = create_translate(input_ndim, -(output_shape - 1) / 2)
+        shift_1 = create_translate(input_ndim, (-(output_shape - 1) / 2).tolist())
         transform = shift @ transform @ shift_1
 
         xform = AffineTransform(
@@ -985,6 +985,7 @@ def __call__(
             else:
                 raise ValueError("Incompatible values: grid=None and spatial_size=None.")
 
+        affine: Union[torch.Tensor, np.ndarray]
         if self.affine is None:
             spatial_dims = len(grid.shape) - 1
             affine = np.eye(spatial_dims + 1)
@@ -1138,7 +1139,7 @@ def __init__(
 
         self.rand_mag = 1.0
         self.as_tensor_output = as_tensor_output
-        self.random_offset = 0.0
+        self.random_offset: np.ndarray
         self.device = device
 
     def randomize(self, grid_size: Sequence[int]) -> None:
@@ -1694,7 +1695,7 @@ def __init__(
         self.padding_mode: GridSamplePadMode = GridSamplePadMode(padding_mode)
         self.device = device
 
-        self.rand_offset = None
+        self.rand_offset: np.ndarray
         self.magnitude = 1.0
         self.sigma = 1.0
 

monai/transforms/spatial/dictionary.py

@@ -1521,7 +1521,7 @@ def inverse(self, data: Mapping[Hashable, np.ndarray]) -> Dict[Hashable, np.ndar
             transform = self.get_most_recent_transform(d, key)
             # Create inverse transform
             zoom = np.array(self.zoomer.zoom)
-            inverse_transform = Zoom(zoom=1 / zoom, keep_size=self.zoomer.keep_size)
+            inverse_transform = Zoom(zoom=(1 / zoom).tolist(), keep_size=self.zoomer.keep_size)
             mode = transform[InverseKeys.EXTRA_INFO]["mode"]
             padding_mode = transform[InverseKeys.EXTRA_INFO]["padding_mode"]
             align_corners = transform[InverseKeys.EXTRA_INFO]["align_corners"]
@@ -1649,7 +1649,7 @@ def inverse(self, data: Mapping[Hashable, np.ndarray]) -> Dict[Hashable, np.ndar
                 mode = transform[InverseKeys.EXTRA_INFO]["mode"]
                 padding_mode = transform[InverseKeys.EXTRA_INFO]["padding_mode"]
                 align_corners = transform[InverseKeys.EXTRA_INFO]["align_corners"]
-                inverse_transform = Zoom(zoom=1 / zoom, keep_size=self.keep_size)
+                inverse_transform = Zoom(zoom=(1 / zoom).tolist(), keep_size=self.keep_size)
                 # Apply inverse
                 d[key] = inverse_transform(
                     d[key],

monai/transforms/utils.py

@@ -224,8 +224,8 @@ def resize_center(img: np.ndarray, *resize_dims: Optional[int], fill_value: floa
 
     resize_dims = fall_back_tuple(resize_dims, img.shape)
 
-    half_img_shape = np.asarray(img.shape) // 2
-    half_dest_shape = np.asarray(resize_dims) // 2
+    half_img_shape = (np.asarray(img.shape) // 2).tolist()
+    half_dest_shape = (np.asarray(resize_dims) // 2).tolist()
     srcslices, destslices = copypaste_arrays(img.shape, resize_dims, half_img_shape, half_dest_shape, resize_dims)
 
     if not inplace:
@@ -318,7 +318,7 @@ def generate_pos_neg_label_crop_centers(
     label_spatial_shape: Sequence[int],
     fg_indices: np.ndarray,
     bg_indices: np.ndarray,
-    rand_state: np.random.RandomState = np.random,
+    rand_state: Optional[np.random.RandomState] = None,
 ) -> List[List[np.ndarray]]:
     """
     Generate valid sample locations based on the label with option for specifying foreground ratio
@@ -338,6 +338,8 @@ def generate_pos_neg_label_crop_centers(
         ValueError: When the foreground and background indices lengths are 0.
 
     """
+    if rand_state is None:
+        rand_state = np.random.random.__self__  # type: ignore
     spatial_size = fall_back_tuple(spatial_size, default=label_spatial_shape)
     if not (np.subtract(label_spatial_shape, spatial_size) >= 0).all():
         raise ValueError("The size of the proposed random crop ROI is larger than the image size.")
@@ -602,7 +604,7 @@ def get_largest_connected_component_mask(img: torch.Tensor, connectivity: Option
 
 
 def get_extreme_points(
-    img: np.ndarray, rand_state: np.random.RandomState = np.random, background: int = 0, pert: float = 0.0
+    img: np.ndarray, rand_state: Optional[np.random.RandomState] = None, background: int = 0, pert: float = 0.0
 ) -> List[Tuple[int, ...]]:
     """
     Generate extreme points from an image. These are used to generate initial segmentation
@@ -624,6 +626,8 @@ def get_extreme_points(
     Raises:
         ValueError: When the input image does not have any foreground pixel.
     """
+    if rand_state is None:
+        rand_state = np.random.random.__self__  # type: ignore
     indices = np.where(img != background)
     if np.size(indices[0]) == 0:
         raise ValueError("get_extreme_points: no foreground object in mask!")

tests/utils.py

@@ -158,7 +158,7 @@ def make_nifti_image(array, affine=None):
 
 def make_rand_affine(ndim: int = 3, random_state: Optional[np.random.RandomState] = None):
     """Create random affine transformation (with values == -1, 0 or 1)."""
-    rs = np.random if random_state is None else random_state
+    rs = np.random.random.__self__ if random_state is None else random_state  # type: ignore
 
     vals = rs.choice([-1, 1], size=ndim)
     positions = rs.choice(range(ndim), size=ndim, replace=False)