def test_stat_aibolit_pipeline(self):
model = self.__load_mock_model()
test_df = generate_fake_dataset()
table = Stats.aibolit_stat(test_df, model)
test_csv = Path(self.cur_file_dir, 'results_test.csv')
results_df = pd.read_csv(test_csv, index_col=0)
all_elements_compared: pd.DataFrame = table.eq(results_df)
bool_eq_elems = np.ravel(all_elements_compared.values)
are_equal_arrays = np.logical_and.reduce(bool_eq_elems, axis=0)
self.assertTrue(are_equal_arrays)
def test_train_with_selected_features():
cur_file_dir = Path(os.path.realpath(__file__)).parent
model = PatternRankingModel()
selected_patterns = ['P18', 'P9', 'M2', 'M5']
train_df = generate_fake_dataset()
print('Features for the whole dataset: {}'.format(list(train_df.columns)))
target = train_df.pop('M4')
start = time()
print('Start training...')
model.fit_regressor(train_df, target, selected_patterns)
end = time()
print('End training. Elapsed time: {:.2f} secs'.format(end - start))
# this folder is created by catboost library, impossible to get rid of it
catboost_folder = Path(cur_file_dir, 'catboost_info')
if catboost_folder.exists():
shutil.rmtree(catboost_folder)
print('Model features: {}'.format(model.features_conf['features_order']))
def test_model_training():
cur_file_dir = Path(os.path.realpath(__file__)).parent
config = Config.get_patterns_config()
model = PatternRankingModel()
patterns = [x['code'] for x in config['patterns']]
train_df = generate_fake_dataset()
model.features_conf = {'features_order': patterns}
scaled_df = scale_dataset(train_df, model.features_conf)
start = time()
print('Start training...')
model.fit_regressor(scaled_df[patterns], scaled_df['M4'])
end = time()
print('End training. Elapsed time: {:.2f} secs'.format(end - start))
# this folder is created by catboost library, impossible to get rid of it
catboost_folder = Path(cur_file_dir, 'catboost_info')
if catboost_folder.exists():
shutil.rmtree(catboost_folder)