len(TypedList) cannot be symbolic

Author: ricardoV94

pytensor/typed_list/basic.py

@@ -12,12 +12,14 @@
 
 
 class _typed_list_py_operators:
+    def __len__(self):
+        raise TypeError(
+            "Cannot call len(TypedList). Use `.length()` method instead for the symbolic equivalent."
+        )
+
     def __getitem__(self, index):
         return getitem(self, index)
 
-    def __len__(self):
-        return length(self)
-
     def append(self, toAppend):
         return append(self, toAppend)
 
@@ -36,10 +38,13 @@ def reverse(self):
     def count(self, elem):
         return count(self, elem)
 
-    # name "index" is already used by an attribute
+    # name "index" is claimed as an attribute of PyTensor Variable(s)
     def ind(self, elem):
         return index_(self, elem)
 
+    def length(self):
+        return length(self)
+
     ttype = property(lambda self: self.type.ttype)
     dtype = property(lambda self: self.type.ttype.dtype)
     ndim = property(lambda self: self.type.ttype.ndim + 1)

tests/typed_list/test_basic.py

@@ -1,3 +1,5 @@
+import re
+
 import numpy as np
 import pytest
 import scipy
@@ -551,8 +553,11 @@ def test_interface(self):
         mySymbolicMatricesList = TypedListType(
             TensorType(pytensor.config.floatX, shape=(None, None))
         )()
-        z = mySymbolicMatricesList.__len__()
-
+        with pytest.raises(
+            TypeError, match=re.escape("Use `.length()` method instead")
+        ):
+            len(mySymbolicMatricesList)
+        z = mySymbolicMatricesList.length()
         f = pytensor.function([mySymbolicMatricesList], z)
 
         x = rand_ranged_matrix(-1000, 1000, [100, 101])