def test_should_include_dataset_subset_name_in_output_file_list(self):
assert get_conf_for_experiment_data({
**DEFAULT_EXPERIMENT_DATA, 'dataset': {
**DEFAULT_EXPERIMENT_DATA['dataset'], 'subset_name':
DATASET_SUBSET_NAME_1
}
})['output_file_list'] == (
f'file-list-{DATASET_SUBSET_NAME_1}.lst')
def test_should_include_eval_name_in_run_name(self):
assert get_conf_for_experiment_data({
**DEFAULT_EXPERIMENT_DATA, 'dataset': {
**DEFAULT_EXPERIMENT_DATA['dataset'], 'eval_name':
DATASET_EVAL_NAME_1
}
})['run_name'] == (
f'{DEFAULT_DATASET_NAME}-{DATASET_EVAL_NAME_1}_{DEFAULT_MODEL_NAME}'
)
def test_should_determine_autocut_train_input_source_file_list(self):
input_model = AUTOCUT_TRAIN_CONFIG_1['autocut']['input_model']
input_model_name = input_model['name']
assert get_conf_for_experiment_data({
**DEFAULT_EXPERIMENT_DATA, 'train':
AUTOCUT_TRAIN_CONFIG_1
})['train']['autocut']['input_source'].get('file_list') == (
f'{DEFAULT_SOURCE_DATA_PATH}-results/{NAMESPACE_1}/'
f'{input_model_name}/file-list.lst')
def test_should_include_eval_name_in_eval_output_path(self):
assert get_conf_for_experiment_data({
**DEFAULT_EXPERIMENT_DATA, 'dataset': {
**DEFAULT_EXPERIMENT_DATA['dataset'], 'eval_name':
DATASET_EVAL_NAME_1
}
})['eval_output_path'] == (
f'{DEFAULT_SOURCE_DATA_PATH}-results/{NAMESPACE_1}/'
f'{DEFAULT_MODEL_NAME}/evaluation-results/{DATASET_EVAL_NAME_1}'
)
def test_should_include_image_config(self):
assert get_conf_for_experiment_data({
**DEFAULT_EXPERIMENT_DATA, 'xyz_image':
IMAGE_1
})['xyz_image'] == IMAGE_1
def test_should_include_config_property(self):
assert get_conf_for_experiment_data({
**DEFAULT_EXPERIMENT_DATA, 'config':
CUSTOM_CONFIG_1
})['config'] == CUSTOM_CONFIG_1
def test_should_include_train_config(self):
assert get_conf_for_experiment_data({
**DEFAULT_EXPERIMENT_DATA, 'train':
TRAIN_CONFIG_1
})['train'] == TRAIN_CONFIG_1
def test_should_enable_resume_by_default(self):
assert get_conf_for_experiment_data({ # pylint: disable=singleton-comparison
**DEFAULT_EXPERIMENT_DATA
})['resume'] is True
def test_should_disable_resume_if_disabled_by_experiment(self):
assert get_conf_for_experiment_data({ # pylint: disable=singleton-comparison
**DEFAULT_EXPERIMENT_DATA, 'resume':
'false'
})['resume'] is False
def test_should_fallback_to_environ_default_limit(
self, get_default_limit_mock):
get_default_limit_mock.return_value = '123'
assert get_conf_for_experiment_data({**DEFAULT_EXPERIMENT_DATA
})['limit'] == '123'
def test_should_determine_source_file_list(self):
assert get_conf_for_experiment_data(
{**DEFAULT_EXPERIMENT_DATA
})['source_file_list'] == DEFAULT_FILE_LIST
def test_should_raise_exception_with_unknown_config(self):
with pytest.raises(InvalidConfigError):
get_conf_for_experiment_data({
**DEFAULT_EXPERIMENT_DATA, 'train':
UNKNOWN_TRAIN_CONFIG_1
})
def test_should_use_default_to_grobid_train_tasks_with_grobid_train_config(
self):
assert get_conf_for_experiment_data({
**DEFAULT_EXPERIMENT_DATA, 'train':
GROBID_TRAIN_CONFIG_1
})['tasks'] == DEFAULT_GROBID_TRAIN_TASKS
def test_should_determine_run_name(self):
assert get_conf_for_experiment_data(
{**DEFAULT_EXPERIMENT_DATA})['run_name'] == (
f'{DEFAULT_DATASET_NAME}_{DEFAULT_MODEL_NAME}')
def test_should_determine_eval_output_path(self):
assert get_conf_for_experiment_data(
{**DEFAULT_EXPERIMENT_DATA})['eval_output_path'] == (
f'{DEFAULT_SOURCE_DATA_PATH}-results/{NAMESPACE_1}/'
f'{DEFAULT_MODEL_NAME}/evaluation-results/all')
def test_should_determine_output_suffix(self):
assert get_conf_for_experiment_data(
{**DEFAULT_EXPERIMENT_DATA})['output_suffix'] == '.xml.gz'
def test_should_include_limit_in_output_file_list(self):
assert get_conf_for_experiment_data({
**DEFAULT_EXPERIMENT_DATA, 'limit':
f'{LIMIT_1}'
})['output_file_list'] == (f'file-list-{LIMIT_1}.lst')
def test_should_include_tasks(self):
tasks = ['task1', 'task2']
assert get_conf_for_experiment_data({
**DEFAULT_EXPERIMENT_DATA, 'tasks':
tasks
})['tasks'] == tasks
def test_should_determine_source_data_path(self):
assert get_conf_for_experiment_data(
{**DEFAULT_EXPERIMENT_DATA
})['source_data_path'] == DEFAULT_SOURCE_DATA_PATH
def test_should_use_default_to_convert_tasks_without_train_config(
self):
assert get_conf_for_experiment_data(
{**DEFAULT_EXPERIMENT_DATA
})['tasks'] == DEFAULT_CONVERT_AND_EVALUATE_TASKS
def test_should_determine_output_file_list(self):
assert get_conf_for_experiment_data(
{**DEFAULT_EXPERIMENT_DATA
})['output_file_list'] == ('file-list.lst')
def test_should_use_default_to_autocut_train_tasks_with_autocut_train_config(
self):
assert get_conf_for_experiment_data({
**DEFAULT_EXPERIMENT_DATA, 'train':
AUTOCUT_TRAIN_CONFIG_1
})['tasks'] == DEFAULT_AUTOCUT_TRAIN_TASKS
def test_should_use_limit_from_experiment(self):
assert get_conf_for_experiment_data({
**DEFAULT_EXPERIMENT_DATA, 'limit':
'123'
})['limit'] == '123'
def test_should_fallback_to_dataset_limit(self):
assert get_conf_for_experiment_data({
**DEFAULT_EXPERIMENT_DATA, 'dataset': {
**DEFAULT_EXPERIMENT_DATA['dataset'], 'limit': '123'
}
})['limit'] == '123'
def test_should_pass_through_model(self):
assert get_conf_for_experiment_data(
{**DEFAULT_EXPERIMENT_DATA
})['model'] == DEFAULT_EXPERIMENT_DATA['model']
