Add bad_indices_policy to TF Probability wrappers.

PiperOrigin-RevId: 655007060

tensorflow_probability/python/internal/backend/numpy/misc.py

@@ -118,11 +118,13 @@ def _sort(values, axis=-1, direction='ASCENDING', name=None):  # pylint: disable
 
 
 # TODO(b/140685491): Add unit-test.
-def _tensor_scatter_nd_add(tensor, indices, updates, name=None):  # pylint: disable=unused-argument
+def _tensor_scatter_nd_add(
+    tensor, indices, updates, bad_indices_policy='', name=None):  # pylint: disable=unused-argument
   """Numpy implementation of `tf.tensor_scatter_nd_add`."""
   indices = _convert_to_tensor(indices)
   tensor = _convert_to_tensor(tensor)
   updates = _convert_to_tensor(updates)
+  del bad_indices_policy
   indices = tuple(
       indices[..., i] for i in range(indices.shape[-1]))  # TODO(b/140685491)
   if JAX_MODE:
@@ -132,11 +134,13 @@ def _tensor_scatter_nd_add(tensor, indices, updates, name=None):  # pylint: disa
 
 
 # TODO(b/140685491): Add unit-test.
-def _tensor_scatter_nd_sub(tensor, indices, updates, name=None):  # pylint: disable=unused-argument
+def _tensor_scatter_nd_sub(
+    tensor, indices, updates, bad_indices_policy='', name=None):  # pylint: disable=unused-argument
   """Numpy implementation of `tf.tensor_scatter_nd_sub`."""
   indices = _convert_to_tensor(indices)
   tensor = _convert_to_tensor(tensor)
   updates = _convert_to_tensor(updates)
+  del bad_indices_policy
   indices = tuple(
       indices[..., i] for i in range(indices.shape[-1]))  # TODO(b/140685491)
   if JAX_MODE:
@@ -146,11 +150,13 @@ def _tensor_scatter_nd_sub(tensor, indices, updates, name=None):  # pylint: disa
 
 
 # TODO(b/140685491): Add unit-test.
-def _tensor_scatter_nd_update(tensor, indices, updates, name=None):  # pylint: disable=unused-argument
+def _tensor_scatter_nd_update(
+    tensor, indices, updates, bad_indices_policy='', name=None):  # pylint: disable=unused-argument
   """Numpy implementation of `tf.tensor_scatter_nd_update`."""
   indices = _convert_to_tensor(indices)
   tensor = _convert_to_tensor(tensor)
   updates = _convert_to_tensor(updates)
+  del bad_indices_policy
   indices = tuple(
       indices[..., i] for i in range(indices.shape[-1]))  # TODO(b/140685491)
   if JAX_MODE: