def test_generate_statistics_some_codes(self):
needed_feature_codes = random.sample(self.feature_codes, k=2)
file_processor = FileProcessor(self.input_file_path)
actual_stats = file_processor.generate_statistics(needed_feature_codes)
_, feature_codes, features = read_test_data_as_arrays(
self.input_file_path)
selected_indexes = []
for code in needed_feature_codes:
indexes = np.where(feature_codes == code)[0].tolist()
selected_indexes += indexes
selected_features = features[selected_indexes]
expected_stats = {
"count": np.uint32(selected_features.shape[0]),
"mean": selected_features.mean(axis=0),
"std": selected_features.std(axis=0, ddof=1),
"max": np.amax(selected_features, axis=0),
"min": np.amin(selected_features, axis=0)
}
for metric in expected_stats:
np.testing.assert_allclose(actual_stats[metric],
expected_stats[metric])
def test_generate_statistics_all_codes(self):
file_processor = FileProcessor(self.input_file_path)
actual_stats = file_processor.generate_statistics()
_, _, features = read_test_data_as_arrays(self.input_file_path)
expected_stats = {
"count": np.uint32(features.shape[0]),
"mean": features.mean(axis=0),
"std": features.std(axis=0, ddof=1),
"max": np.amax(features, axis=0),
"min": np.amin(features, axis=0)
}
for metric in expected_stats:
np.testing.assert_allclose(actual_stats[metric],
expected_stats[metric])