stepsel.binning.helper
======================

.. py:module:: stepsel.binning.helper

.. autoapi-nested-parse::

   Helper functions for binning. Smaller functions that are used in multiple binning functions.



Functions
---------

.. autoapisummary::

   stepsel.binning.helper.get_tree_cut_points
   stepsel.binning.helper.bin_values


Module Contents
---------------

.. py:function:: get_tree_cut_points(clf: sklearn.tree.DecisionTreeRegressor | sklearn.tree.DecisionTreeClassifier, feature_names: numpy.typing.ArrayLike | None = None)

   Get the cut points of a decision tree.

   :param clf: The decision tree to get the cut points from.
   :type clf: DecisionTreeRegressor or DecisionTreeClassifier
   :param feature_names: The feature names of the decision tree. If None, the features are assumed to be integers.
   :type feature_names: array-like, optional

   :returns: **feature_cut_points** -- A dictionary with the feature names as keys and the cut points as values.
   :rtype: dict


.. py:function:: bin_values(data: numpy.typing.ArrayLike, thresholds: numpy.typing.ArrayLike, right=True) -> numpy.ndarray

   Bin data into bins based on thresholds.

   :param data: The input values to be binned.
   :type data: array-like
   :param thresholds: The thresholds to use for binning, ordered from smallest to largest.
   :type thresholds: array-like
   :param right: Whether the intervals should be closed on the right (default) or left.
   :type right: bool, optional

   :returns: **binned_values** -- The binned values. String format is "(a, b]" if right=True, "[a, b)" if right=False.
   :rtype: array-like

   .. todo:: * Add option to return pd.Categorical ordered by thresholds.