[pyspark] Validate the validation indicator column type.

[trivialfis]

Close #11496 .

[trivialfis]

cc @wbo4958 @WeichenXu123

[Copilot]

Pull Request Overview

This pull request introduces validation to ensure that the validation indicator column is of boolean type. Key changes include:

• Addition of a new test (test_valid_type) in the Spark test suite to verify that non-boolean indicator values result in an error.
• Introduction of the helper function is_bool_column in the Spark data module.
• Update of the make_blob function to raise a TypeError when the validation indicator column is not boolean.

Reviewed Changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.

File	Description
tests/test_distributed/test_with_spark/test_spark_local.py	Added a negative test case to confirm the validation indicator column type check
python-package/xgboost/spark/data.py	Added is_bool_column helper and updated make_blob to raise an error if the column type is invalid

Comments suppressed due to low confidence (2)

tests/test_distributed/test_with_spark/test_spark_local.py:1311

• [nitpick] Consider adding a brief inline comment to explain why a generic Exception is expected in this test case due to the variability of Spark's error types.

    def test_valid_type(self, spark: SparkSession) -> None:

python-package/xgboost/spark/data.py:75

• [nitpick] It may be helpful to include an inline comment clarifying the dual check using both cuDF and pandas boolean type verifications to improve maintainability.

            if not is_bool_column(col.dtype):