[BUG] Groupby-apply returns DataFrame rather than Series when taking the JIT path

[shwina]

Depending on whether I choose engine='jit' or engine='cudf', I get back a DataFrame rather than the expected Series:

In [3]: df = cudf.DataFrame({'a': [1, 1, 1, 2, 2], 'b': [1, 2, 3, 4, 5]})

In [4]: df.groupby('a').apply(lambda group: group['a'].sum())
Out[4]:
   a  None
0  1     3
1  2     4

In [5]: df.groupby('a').apply(lambda group: group['a'].sum(), engine='cudf')
Out[5]:
a
1    3
2    4
dtype: int64