Construct tests for model_tuning_core()

[ETA444]

Test Descriptions

1. Type Error Tests:

   • Invalid x_train Type: Ensures a TypeError is raised when x_train is neither a DataFrame nor an ndarray.
   • Invalid y_train Type: Checks for a TypeError if y_train is not a Series or ndarray.
   • Invalid task_type: Verifies that providing a non-string task_type raises a TypeError, and an incorrect string raises a ValueError.
   • Invalid models Type: Asserts a TypeError is raised if models is not a dictionary.
   • Invalid priority_metrics Type: Checks for a TypeError when priority_metrics is not a list.
   • Invalid priority_tuners Type: Verifies a TypeError for non-list priority_tuners.
   • Invalid custom_param_grids Type: Asserts a TypeError if custom_param_grids is not a dictionary.
   • Invalid n_jobs, cv, verbose, and random_state Types: Ensures TypeErrors are raised if these parameters are not integers.
   • Invalid cv Value: Confirms a ValueError if cv is less than 1.
   • Invalid Iteration Values: Checks for ValueErrors when n_iter_random or n_iter_bayesian are less than 1.

2. Value Error Tests:

   • Mismatched x_train and y_train Shapes: Ensures a ValueError is raised if x_train and y_train do not have compatible shapes.
   • Empty x_train or y_train: Confirms ValueErrors are raised if x_train or y_train are empty.
   • Duplicate or Invalid priority_metrics: Verifies that ValueErrors are raised for duplicate values or invalid metric names in priority_metrics.
   • Incompatible Metrics for Task Type: Checks for ValueErrors if metrics not suitable for the specified task_type are provided.
   • Excessive n_top_models: Ensures a ValueError if n_top_models exceeds the number of available models for the specified task.
   • Unrecognized priority_tuners: Verifies a ValueError for unrecognized tuner names.
   • Inappropriate refit_metric: Checks for a ValueError if refit_metric is not suitable for the given task_type.

3. Functionality tests: ... *

Example Test Code

Here's an example of testing invalid x_train type:

def test_model_tuning_core_invalid_x_train_type(sample_data_mrc_mtc):
    """
    Tests if model_tuning_core raises TypeError when 'x_train' is not a DataFrame or ndarray.
    """
    _, y_train_classification, _ = sample_data_mrc_mtc
    with pytest.raises(TypeError):
        model_tuning_core(x_train="invalid_type", y_train=y_train_classification, task_type='classification', models={'model': LogisticRegression()})

This test ensures that the function properly checks x_train's data type and raises an appropriate error if it's not a DataFrame or ndarray.

Access the Full Test Suite*

For a comprehensive view of all tests, including their implementations and additional configurations, you can access the full test suite here.