Nick Becker

Add Groupby.rank for DataFrame and Series GroupBy

10

This PR implements `{Series, DataFrame}.groupby.rank` based on the same hash partition pattern used by `groupby.{shift, transform, apply}`. - [x] Closes #8658 - [x] Tests added / passed (locally) - [x]...

For pandas API compatibility, we can implement [Series](https://pandas.pydata.org/docs/reference/api/pandas.Series.truncate.html) and [DataFrame.truncate](https://pandas.pydata.org/docs/reference/api/pandas.DataFrame.truncate.html). Truncate is "a useful shorthand for boolean indexing based on index values above or below certain thresholds." The DataFrame method...

[DOC] Update FIL docs to include support for categorical LightGBM

1

As we now support categorical LightGBM models , we should update the FIL docs - https://github.com/rapidsai/cuml/blob/d78a8056511788f0d04554e376a4c075eb7135c6/python/cuml/fil/fil.pyx#L489 - https://github.com/rapidsai/cuml/tree/branch-22.08/python/cuml/fil#features Context: https://github.com/rapidsai/cuml/issues/1424#issuecomment-1170230052

[FEA] cuML estimators should support non-consecutive labels outside of [0, n) where appropriate

9

As noted in https://github.com/rapidsai/cudf/issues/10024 , cuML RandomForestClassifier will throw an error if the target column has non-consecutive labels outside of the [0, n) range. This does not occur in scikit-learn,...

[FEA] Support aliases in aggregation clauses

4

I'd like to be able to do a GROUP BY aggregation and then filter the resulting aggregated data based on a condition using the HAVING clause while including the aggregation...

[FEA] Support LCASE alias for LOWER

2

I'd like to be able to call LCASE on a string column to convert it to lowercase, like in MySQL. This is an alias for LOWER, which is noted in...

GPU backend dependencies aren't included in the development environment yml files, causing pytests to fail with `--rungpu` out of the box. Datafusion branch: https://github.com/dask-contrib/dask-sql/blob/datafusion-sql-planner/continuous_integration/environment-3.10-dev.yaml It would be useful for there...

[FEA] Support for HDBSCAN membership_vector and all_points_membership_vectors

6

I'd like to be able to use HDBSCAN to calculate membership vectors for points, like I can with the CPU library. Per the CPU library [documentation](https://hdbscan.readthedocs.io/en/latest/api.html#hdbscan.prediction.membership_vector), this function "produces a...

Update KMeans notebook for clarity

1

This PR updates the existing KMeans notebook to improve clarity and make the performance benefit clear. Closes https://github.com/rapidsai/cuml/issues/4135

For API compatibility, I'd like to be able to generate prediction data after the fact using `generate_prediction_data`, even if I set `prediction_data=False` when I instantiated my clusterer. I can do...