def test_get_skip_evaluation_pipeline_and_parameters(self):
_, parameter_values = utils.get_skip_evaluation_pipeline_and_parameters(
'project', 'us-central1', 'gs://foo', 'target', 'classification',
'maximize-au-prc', {'auto': {
'column_name': 'feature_1'
}}, {
'fraction_split': {
'training_fraction': 0.8,
'validation_fraction': 0.2,
'test_fraction': 0.0
}
}, {'csv_data_source': {
'csv_filenames': ['gs://foo/bar.csv']
}}, 1000)
expected_parameter_values = {
'project': 'project',
'location': 'us-central1',
'root_dir': 'gs://foo',
'target_column_name': 'target',
'prediction_type': 'classification',
'optimization_objective': 'maximize-au-prc',
'transformations':
'{\\"auto\\": {\\"column_name\\": \\"feature_1\\"}}',
'split_spec':
'{\\"fraction_split\\": {\\"training_fraction\\": 0.8, '
'\\"validation_fraction\\": 0.2, \\"test_fraction\\": 0.0}}',
'data_source': '{\\"csv_data_source\\": {\\"csv_filenames\\": '
'[\\"gs://foo/bar.csv\\"]}}',
'stage_1_deadline_hours': 0.7708333333333334,
'stage_1_num_parallel_trials': 35,
'stage_1_num_selected_trials': 7,
'stage_1_single_run_max_secs': 634,
'reduce_search_space_mode': 'minimal',
'stage_2_deadline_hours': 0.22916666666666663,
'stage_2_num_parallel_trials': 35,
'stage_2_num_selected_trials': 5,
'stage_2_single_run_max_secs': 634,
'weight_column_name': '',
'optimization_objective_recall_value': -1,
'optimization_objective_precision_value': -1,
'study_spec_override': '',
'stage_1_tuner_worker_pool_specs_override': '',
'cv_trainer_worker_pool_specs_override': '',
'export_additional_model_without_custom_ops': False,
'stats_and_example_gen_dataflow_machine_type': 'n1-standard-16',
'stats_and_example_gen_dataflow_max_num_workers': 25,
'stats_and_example_gen_dataflow_disk_size_gb': 40,
'transform_dataflow_machine_type': 'n1-standard-16',
'transform_dataflow_max_num_workers': 25,
'transform_dataflow_disk_size_gb': 40,
'encryption_spec_key_name': '',
'dataflow_subnetwork': '',
'dataflow_use_public_ips': True,
}
self.assertEqual(parameter_values, expected_parameter_values)
def test_get_skip_evaluation_pipeline_and_parameters(self):
_, parameter_values = utils.get_skip_evaluation_pipeline_and_parameters(
'project', 'us-central1', 'gs://foo', 'target', 'classification',
'maximize-au-prc', {'auto': {
'column_name': 'feature_1'
}}, {
'fraction_split': {
'training_fraction': 0.8,
'validation_fraction': 0.2,
'test_fraction': 0.0
}
}, {'csv_data_source': {
'csv_filenames': ['gs://foo/bar.csv']
}}, 1000)
self.assertEqual(parameter_values, self.parameter_values)
