Fix unbounded sequence issue in DataFrame constructor (#13811)

In `cudf`, we currently have a hang in this scenario:
```python
In [1]: import cudf

In [2]:     class A:
   ...:         def __getitem__(self, key):
   ...:             return 1
   ...: 

In [3]: cudf.DataFrame([A()])
```

This PR introduces additional checks before letting the list-like inputs pass onto `itertools` for transposing.

Authors:
  - GALI PREM SAGAR (https://github.com/galipremsagar)

Approvers:
  - Bradley Dice (https://github.com/bdice)

URL: #13811

python/cudf/cudf/core/dataframe.py

@@ -843,6 +843,12 @@ def _init_from_list_like(self, data, index=None, columns=None):
             data = DataFrame.from_pandas(pd.DataFrame(data))
             self._data = data._data
         else:
+            if any(
+                not isinstance(col, (abc.Iterable, abc.Sequence))
+                for col in data
+            ):
+                raise TypeError("Inputs should be an iterable or sequence.")
+
             data = list(itertools.zip_longest(*data))
 
             if columns is not None and len(data) == 0:

python/cudf/cudf/tests/test_dataframe.py

@@ -10243,3 +10243,15 @@ def test_dataframe_init_columns_named_index():
     pdf = pd.DataFrame(data, columns=columns)
 
     assert_eq(gdf, pdf)
+
+
+def test_dataframe_constructor_unbounded_sequence():
+    class A:
+        def __getitem__(self, key):
+            return 1
+
+    with pytest.raises(TypeError):
+        cudf.DataFrame([A()])
+
+    with pytest.raises(TypeError):
+        cudf.DataFrame({"a": A()})