tst.py

def test_all_Steps_of_pipeline(
        dummy_pipeline: PipelineTuple, expected_dronology_ypred_ytrue
):
    """
    Test loading predictions from a CSV file into the pipeline and calculate
    the metrics and export the results.

    We aim to validate that the pipeline can successfully load prediction results from a CSV
    file. The metrics are tested through the exporter, because they calculate the metrics by
    using the methods from the Metrics class.
    """
    # Load Prediction
    pipeline, _, _ = dummy_pipeline
    pipeline.load_prediction_from_csv(
        Path(
            f"{dir_path}/csvs/NoRBERT_Task4_IsFunctional_e10_NoSampling_Dronology-IsFunctional.csv"
        )
    )

    # Assert Prediction
    y_pred, y_true = expected_dronology_ypred_ytrue
    assert pipeline._prediction_results[0].result.y_pred[:10] == y_pred
    assert pipeline._prediction_results[0].result.y_true[:10] == y_true

    # Assert Metrics and test the exporter
    metrics = Metrics(pipeline._prediction_results)

    with tempfile.TemporaryDirectory() as tmp_dir:
        temp_dir_path = Path(tmp_dir)
        latex_output_path = temp_dir_path.joinpath('metrics.tex')
        csv_output_path = temp_dir_path.joinpath('metrics.csv')
        expected_exporter_output_path = Path(
            f'{os.path.dirname(__file__)}/expected_exporter_output')

        # Latex Exporter
        LatexExporter().export_all_metrics(metrics, str(latex_output_path), None)
        expected_latex_content = expected_exporter_output_path.joinpath('metrics.tex').read_text()
        actual_latex_content = latex_output_path.read_text()
        assert actual_latex_content == expected_latex_content

        # CSV Exporter
        CsvExporter().export_all_metrics(metrics, str(csv_output_path), None)
        expected_csv_content = expected_exporter_output_path.joinpath('metrics.csv').read_text()
        actual_csv_content = csv_output_path.read_text()
        assert actual_csv_content == expected_csv_content