Fix call sites invoking to_type and TensorType using a `tensorflo…

…w` type.

PiperOrigin-RevId: 597676178

tensorflow_federated/python/core/impl/types/computation_types.py

@@ -302,14 +302,12 @@ def _clear_intern_pool() -> None:
 ]
 
 
-def _is_dtype_like(
-    obj: object,
-) -> TypeGuard[Union[_DtypeLike, tf.dtypes.DType]]:
+def _is_dtype_like(obj: object) -> TypeGuard[_DtypeLike]:
   """Returns `True` if `obj` is dtype like, otherwise `False`."""
   if isinstance(obj, type) and issubclass(obj, np.generic):
     return True
   else:
-    return isinstance(obj, (tf.dtypes.DType, np.dtype))
+    return isinstance(obj, np.dtype)
 
 
 _ALLOWED_NP_DTYPES = [
@@ -379,27 +377,18 @@ class TensorType(Type, metaclass=_Intern):
   @classmethod
   def _hashable_from_init_args(
       cls,
-      dtype: Union[_DtypeLike, tf.dtypes.DType],
+      dtype: _DtypeLike,
       shape: array_shape._ArrayShapeLike = (),
   ) -> Hashable:
     """Returns hashable `TensorType.__init__` args."""
-    # TODO: b/305743962 - This is only required to convert a `tf.dtypes.DType`
-    # to a `np.dtype`. It should be when `tf.dtypes.DType` can not be passed
-    # into the constructor of the `tff.TensorType`.
-    if isinstance(dtype, tf.dtypes.DType):
-      if dtype.base_dtype == tf.string:
-        dtype = np.str_
-      else:
-        dtype = dtype.base_dtype.as_numpy_dtype
-
     dtype = _to_dtype(dtype)
     if shape is not None:
       shape = tuple(shape)
     return (dtype, shape)
 
   def __init__(
       self,
-      dtype: Union[_DtypeLike, tf.dtypes.DType],
+      dtype: _DtypeLike,
       shape: array_shape._ArrayShapeLike = (),
   ):
     """Constructs a new instance from the given `dtype` and `shape`.
@@ -411,15 +400,6 @@ def __init__(
     Raises:
       TypeError: if arguments are of the wrong types.
     """
-    # TODO: b/305743962 - This is only required to convert a `tf.dtypes.DType`
-    # to a `np.dtype`. It should be when `tf.dtypes.DType` can not be passed
-    # into the constructor of the `tff.TensorType`.
-    if isinstance(dtype, tf.dtypes.DType):
-      if dtype.base_dtype == tf.string:
-        dtype = np.str_
-      else:
-        dtype = dtype.base_dtype.as_numpy_dtype
-
     self._dtype = _to_dtype(dtype)
     if shape is not None:
       shape = tuple(shape)