Work-around for numpy bug in choice with size=()

pytensor/tensor/random/basic.py

@@ -2084,6 +2084,10 @@ class ChoiceWithoutReplacement(RandomVariable):
         batch_ndim = max(batch_ndim, size_ndim)
 
         if batch_ndim == 0:
+            # Numpy choice fails with size=() if a.ndim > 1 is batched
+            # https://github.com/numpy/numpy/issues/26518
+            if core_shape == ():
+                core_shape = None
             return rng.choice(a, p=p, size=core_shape, replace=False)
 
         # Numpy choice doesn't have a concept of batch dims

tests/tensor/random/test_basic.py

@@ -1422,6 +1422,15 @@ def test_choice_samples():
     compare_sample_values(choice, pt.as_tensor_variable([1, 2, 3]), 2, replace=True)
 
 
+def test_choice_scalar_size():
+    np.testing.assert_array_equal(
+        choice([[1, 2, 3]], size=(), replace=True).eval(), [1, 2, 3]
+    )
+    np.testing.assert_array_equal(
+        choice([[1, 2, 3]], size=(), replace=False).eval(), [1, 2, 3]
+    )
+
+
 def test_permutation_samples():
     compare_sample_values(
         permutation,