[FEA] return first n rows of each group, after doing groupby with cudf.

[rnyak]

Is your feature request related to a problem? Please describe.
I'd like to be able to return first n rows of each group, after doing groupby with cudf. This can be done with pdf as follows:

import pandas as pd
pdf = pd.DataFrame({'id6': 4*list(reversed(range(3))), 'v3': 3*list(range(4))})

pdf[['id6', 'v3']].sort_values('v3', ascending=False).groupby(['id6']).head(2)

       id6	v3
3	2	3
7	1	3
11	0	3
2	0	2
6	2	2
10	1	2

But cudf generates:

``DataFrameGroupBy' object has no attribute 'head'` attribute error.

Describe alternatives you've considered
I could not find an alternative solution yet.

[beckernick]

Thanks for the issue! Closing this as a duplicate of #2592 to consolidate discussion.