https://github.com/microsoft/recommenders/blob/efaa3d7742183dee0846877e2dc64977098e1977/reco_utils/recommender/sar/sar_singlenode.py#L232
We may assume that the input data frame has no duplicates, since 90% of the time this has been ensured in prior data prep stages of the data science pipeline.
- Remove the
drop_duplicates() occurrence
- Update doc string to require input DFs contain no duplicates.