Raise error when two dims are not compatible

pymc_marketing/prior.py

@@ -110,7 +110,7 @@ def custom_transform(x):
 
 
 class UnsupportedShapeError(Exception):
-    """Error for when the shape of the hierarchical variable is not supported."""
+    """Error for when the shapes from variables are not compatible."""
 
 
 class UnsupportedDistributionError(Exception):
@@ -169,6 +169,12 @@ def handle_dims(x: pt.TensorLike, dims: Dims, desired_dims: Dims) -> pt.TensorVa
     dims = dims if isinstance(dims, tuple) else (dims,)
     desired_dims = desired_dims if isinstance(desired_dims, tuple) else (desired_dims,)
 
+    if difference := set(dims).difference(desired_dims):
+        raise UnsupportedShapeError(
+            f"Dims {dims} of data are not a subset of the desired dims {desired_dims}. "
+            f"{difference} is missing from the desired dims."
+        )
+
     aligned_dims = np.array(dims)[:, None] == np.array(desired_dims)
 
     missing_dims = aligned_dims.sum(axis=0) == 0

tests/test_prior.py

@@ -72,6 +72,20 @@ def test_handle_dims(x, dims, desired_dims, expected_fn) -> None:
     np.testing.assert_array_equal(result, expected_fn(x))
 
 
+@pytest.mark.parametrize(
+    "x, dims, desired_dims",
+    [
+        (np.ones(3), "channel", "something_else"),
+        (np.ones((3, 2)), ("a", "b"), ("a", "B")),
+    ],
+    ids=["no_incommon", "some_incommon"],
+)
+def test_handle_dims_with_impossible_dims(x, dims, desired_dims) -> None:
+    match = " are not a subset of the desired dims "
+    with pytest.raises(UnsupportedShapeError, match=match):
+        handle_dims(x, dims, desired_dims)
+
+
 def test_missing_transform() -> None:
     match = "Neither pytensor.tensor nor pymc.math have the function 'foo_bar'"
     with pytest.raises(UnknownTransformError, match=match):