Adds parameterized tests for the evaluate function's parallelization backends to ensure its robustness across different parallel execution configurations in Time Series Classification.

What was done

• Modified sktime/classification/model_evaluation/tests/test_evaluate.py to enhance testing for parallel backends.
• Introduced _get_parallel_test_fixtures to retrieve available parallel backends for testing.
• Parameterized the test_evaluate_parallel_backend function using pytest.mark.parametrize to run tests for various backends.
• Updated the evaluate function call within the test to dynamically use the backend and backend_params from the parameterized fixtures.
• Included a minor style fix for mixed line endings in the test file.

Impact

• Improves test coverage and reliability for the evaluate function's parallel execution capabilities.
• Ensures that the evaluate function correctly handles different parallelization backends (e.g., loky, multiprocessing, threading).
• Enhances developer confidence in the stability and robustness of sktime's evaluation utilities, especially concerning parallel processing.

Technical details

• Affected file: sktime/classification/model_evaluation/tests/test_evaluate.py.
• Utilized pytest.mark.parametrize for data-driven testing, allowing the same test to run with multiple parallel backend configurations.
• Integrated sktime.utils.parallel._get_parallel_test_fixtures to abstract and standardize the setup of parallel test fixtures.
• The evaluate function call in the test now dynamically unpacks backend parameters using **backend.
• The changes primarily focus on the testing framework rather than the core evaluate logic itself.

Related issues

• #8703

Metadata

• Source: github
• Repo: sktime/sktime
• URL: https://github.com/sktime/sktime/pull/8737
• Merged: 2025-08-24T12:46:02Z