[FEA]: Support UDF expression for referencing self in assign

[mroeschke]

Is your feature request related to a problem? Please describe.
While working on a pandas to cudf workflow comparison, I noticed that calling assign with a lambda that references the current DataFrame is not supported

Describe the solution you'd like

In [34]: df = pd.DataFrame({"a": [1]})

In [35]: df.assign(b=lambda x: x["a"] + 1)
Out[35]:
   a  b
0  1  2

In [36]: cu_df = cudf.DataFrame({"a": [1]})

In [37]: cu_df.assign(b=lambda x: x["a"] + 1)
TypeError: 'function' object is not iterable

During handling of the above exception, another exception occurred:

ValueError: Unsupported dtype object

During handling of the above exception, another exception occurred:

TypeError: 'function' object is not iterable

Describe alternatives you've considered
Just assigning the column outside the lambda. Unfortunately it's not conducive to method chaining use cases

Additional context
https://pandas.pydata.org/docs/reference/api/pandas.DataFrame.assign.html?highlight=assign#pandas.DataFrame.assign

[wence-]

Probably just

diff --git a/python/cudf/cudf/core/dataframe.py b/python/cudf/cudf/core/dataframe.py
index e50c324a8f..b17ba765f7 100644
--- a/python/cudf/cudf/core/dataframe.py
+++ b/python/cudf/cudf/core/dataframe.py
@@ -1470,14 +1470,21 @@ class DataFrame(IndexedFrame, Serializable, GetAttrGetItemMixin):
         2  2  5
         """
         new_df = cudf.DataFrame(index=self.index.copy())
+
+        def make_col(col):
+            if callable(col):
+                return col(self)
+            else:
+                return col
+
         for name, col in self._data.items():
             if name in kwargs:
-                new_df[name] = kwargs.pop(name)
+                new_df[name] = make_col(kwargs.pop(name))
             else:
                 new_df._data[name] = col.copy()
 
         for k, v in kwargs.items():
-            new_df[k] = v
+            new_df[k] = make_col(v)
         return new_df
 
     @classmethod

Can you give that a try?