XGBoost model trained with features_cols parameter can also predict vector or array feature type. But first we need to check features_cols

Gets the hostIP for Spark. This essentially gets the IP of the first worker.

Wrap log_callback() method in ctypes callback type.

Return the true prediction function which will be running on the executor side

Get rabit context arguments from data distribution in DaskDMatrix.

Try to enable stage-level scheduling

Convert list of dictionaries into a dictionary of lists.

Get the XGBoost version from native shared object.

Load xgboost library.

Load model attributes without hyper-parameters.

Handle numpy slice. This can be removed if we use __array_interface__.

Help transform pandas series for meta data like labels

Post process of transform

Subclasses should override this method and returns a _SparkXGBModel subclass

Reference data from numpy array.

Reference data from a CPU DataFrame.

Reference data from scipy csr.

Reference data from CUDA columnar format.

Sparse pandas columns are not supported with column-wise data split.

We repartition the dataset if the number of workers is not equal to the number of partitions.

If train or transform on the gpu according to the parameters

Get the xgboost.XGBModel().fit() parameters and set them to spark parameters

Get the parameters from xgboost.XGBModel().predict() and set them into spark parameters

Temporarily set the client for sklearn model.

Set xgboost parameters into spark parameters

Start Rabit tracker with n_workers

Update ``sklearn.utils.Tags`` inherited from ``scikit-learn`` base classes. ``scikit-learn`` 1.6 introduced a dataclass-based interface for e

Aggregate cross-validation results. @param results eval info from each data sample @return cross-validation eval info

The approx tree method uses the basic `DMatrix` (not recommended).

Get the array interface representation of a string-based category array.

Get attributes stored in the Booster as a dictionary. Returns ------- result : dictionary of attribute_name: attribute_value

Parameters shared by the three models in the demo.

Run before each iteration. Returns True when training should stop. See :py:meth:`after_iteration` for details.

NOLINT

Boost the booster for one iteration with customized gradient statistics. .. warning:: Like :py:meth:`.update`, this function sho

Build the native library.

Build libxgboost in a temporary directory and obtain the path to built libxgboost.

Train using builtin categorical data support from XGBoost

Make sure no missing value in meta data.

Test for the `expectile` loss.

Test init estimation for classsifier.

Test init estimation for regressor.

Ensure the script is run from the project root directory.

Demo for using the checkpoint callback. Custom logic for handling output is usually required and users are encouraged to define their own callback