Add shape_tuple helper function

aesara/tensor/basic.py

@@ -42,6 +42,7 @@ from aesara.tensor.shape import (
     shape_padaxis,
     shape_padleft,
     shape_padright,
+    shape_tuple,
 )
 from aesara.tensor.type import (
     TensorType,
@@ -4152,7 +4153,7 @@ class Choose(Op):
         else:
             choice = as_tensor_variable(choices)
         (out_shape,) = self.infer_shape(
-            None, None, [tuple(a.shape), tuple(shape(choice))]
+            None, None, [shape_tuple(a), shape_tuple(choice)]
         )
 
         bcast = []

aesara/tensor/random/op.py

@@ -20,6 +20,7 @@ from aesara.tensor.elemwise import Elemwise
 from aesara.tensor.exceptions import NotScalarConstantError
 from aesara.tensor.random.type import RandomType
 from aesara.tensor.random.utils import normalize_size_param, params_broadcast_shapes
+from aesara.tensor.shape import shape_tuple
 from aesara.tensor.type import TensorType, all_dtypes
 
 
@@ -199,7 +200,7 @@ class RandomVariable(Op):
 
         # Broadcast the parameters
         param_shapes = params_broadcast_shapes(
-            param_shapes or [p.shape for p in dist_params], self.ndims_params
+            param_shapes or [shape_tuple(p) for p in dist_params], self.ndims_params
         )
 
         def slice_ind_dims(p, ps, n):

aesara/tensor/shape.py

@@ -129,6 +129,19 @@ shape = Shape()
 _shape = shape  # was used in the past, now use shape directly.
 
 
+def shape_tuple(x):
+    """Get a tuple of symbolic shape values.
+
+    This will return a `ScalarConstant` with the value ``1`` wherever
+    broadcastable is ``True``.
+    """
+    one_at = aesara.scalar.ScalarConstant(aesara.scalar.int64, 1)
+    return tuple(
+        one_at if getattr(sh, "value", sh) == 1 or bcast else sh
+        for sh, bcast in zip(shape(x), getattr(x, "broadcastable", (False,) * x.ndim))
+    )
+
+
 class Shape_i(COp):
     """
     L{Op} to return the shape of a matrix.