Dask Groupby Multi-index Level

I want to groupby dask multi-index data frame by its level. I want to do the following pandas equivalent in dask:

df.groupby(level=0)['TARGET']\
                .apply(lambda x: x.shift().rolling(min_periods=1, window=7).sum()).fillna(0)\
                .reset_index(name=str(7)+"D")

The document says that dask doesn't support pandas level argument. How should I tackle this problem in that case?

I can't use the column name because the column is in the index and the column index which throws the following error:

ValueError: 'CONTACT_ID' is both an index level and a column label, which is ambiguous.

Let me know if you would require data in order to be able to assist. Thanks a lot in advance.