def test_multi_task_prediction_list_to_json(self):
featurized_jsons = glob(
os.path.join(MULTI_TASK_FEATURES_PATH, "*features.json"))
json_obj = {
"mode": self.events_mode,
"file_list": featurized_jsons,
'run_list': list(range(len(featurized_jsons)))
}
json_string = json.dumps(json_obj)
newjsonpaths = process_file_list_from_json(
json_string,
model_dir=MODEL_DIR,
processed_dir=MULTI_TASK_FEATURES_PATH,
predict_only=True)
reloaded = json.loads(newjsonpaths)
prediction_reloaded = loadfn(reloaded['file_list'][0])
self.assertIsInstance(prediction_reloaded['cycle_number'], np.ndarray)
self.assertGreater(len(prediction_reloaded['cycle_number']), 1)
# Testing error output
self.assertIsInstance(prediction_reloaded['fractional_error'],
np.ndarray)
self.assertEqual(len(prediction_reloaded['fractional_error']),
1) # for now just a single fractional error
predictions = glob(
os.path.join(MULTI_TASK_FEATURES_PATH, "*predictions.json"))
for file in predictions:
os.remove(file)
def test_serialized_prediction(self):
# Testing nominal capacity calculation
feature_json_path = os.path.join(
TEST_FILE_DIR,
"2017-06-30_2C-10per_6C_CH10_full_model_multi_features.json")
json_obj = {
"mode": self.events_mode,
"file_list": [feature_json_path, feature_json_path],
'run_list': [0, 1]
}
json_string = json.dumps(json_obj)
newjsonpaths = process_file_list_from_json(json_string,
model_dir=MODEL_DIR,
processed_dir=TEST_FILE_DIR)
reloaded = json.loads(newjsonpaths)
prediction_reloaded = loadfn(reloaded['file_list'][0])
features = loadfn(feature_json_path)
self.assertEqual(features.nominal_capacity, 1.0628421000000001)
self.assertFalse((prediction_reloaded['discharge_capacity'] -
np.around(np.arange(.98, 0.78, -0.03), 2) *
features.nominal_capacity).any())
os.remove(
os.path.join(
TEST_FILE_DIR,
'2017-06-30_2C-10per_6C_CH10_full_model_multi_predictions.json'
))
def test_single_task_prediction_list_to_json(self):
featurized_jsons = glob(
os.path.join(SINGLE_TASK_FEATURES_PATH, "*features.json"))
json_obj = {
"mode": self.events_mode,
"file_list": featurized_jsons,
"run_list": list(range(len(featurized_jsons))),
}
json_string = json.dumps(json_obj)
newjsonpaths = process_file_list_from_json(
json_string,
model_dir=MODEL_DIR,
processed_dir=SINGLE_TASK_FEATURES_PATH,
predict_only=True,
)
reloaded = json.loads(newjsonpaths)
# Ensure first is correct
prediction_reloaded = loadfn(reloaded["file_list"][0])
self.assertIsInstance(prediction_reloaded["cycle_number"], np.ndarray)
# Testing error output
self.assertIsInstance(prediction_reloaded["fractional_error"],
np.ndarray)
predictions = glob(
os.path.join(SINGLE_TASK_FEATURES_PATH, "*predictions.json"))
for file in predictions:
os.remove(file)
def test_single_task_prediction_list_to_json(self):
featurized_jsons = glob(
os.path.join(single_task_features_path, "*features.json"))
json_obj = {
"mode": self.events_mode,
"file_list": featurized_jsons,
'run_list': list(range(len(featurized_jsons)))
}
json_string = json.dumps(json_obj)
newjsonpaths = process_file_list_from_json(
json_string,
model_dir=TEST_FILE_DIR,
processed_dir=single_task_features_path,
predict_only=True)
reloaded = json.loads(newjsonpaths)
# Ensure first is correct
prediction_reloaded = loadfn(reloaded['file_list'][0])
self.assertIsInstance(prediction_reloaded['cycle_number'], np.ndarray)
# Testing error output
self.assertIsInstance(prediction_reloaded['fractional_error'],
np.ndarray)
predictions = glob(
os.path.join(single_task_features_path, "*predictions.json"))
for file in predictions:
os.remove(file)
def test_python(self):
"""Python script for end to end test"""
# Copy
mapped = collate.process_files_json()
rename_output = json.loads(mapped)
rename_output['mode'] = self.events_mode # mode run|test|events_off
rename_output['run_list'] = list(range(len(
rename_output['file_list'])))
mapped = json.dumps(rename_output)
# Validation
validated = validate.validate_file_list_from_json(mapped)
validated_output = json.loads(validated)
validated_output['mode'] = self.events_mode # mode run|test|events_off
validated_output['run_list'] = list(
range(len(validated_output['file_list'])))
validated = json.dumps(validated_output)
# Data structuring
structured = structure.process_file_list_from_json(validated)
structured_output = json.loads(structured)
structured_output[
'mode'] = self.events_mode # mode run|test|events_off
structured_output['run_list'] = list(
range(len(structured_output['file_list'])))
structured = json.dumps(structured_output)
# Featurization
featurized = featurize.process_file_list_from_json(structured)
featurized_output = json.loads(featurized)
featurized_output[
'mode'] = self.events_mode # mode run|test|events_off
featurized_output['run_list'] = list(
range(len(featurized_output['file_list'])))
featurized = json.dumps(featurized_output)
# Prediction
predictions = run_model.process_file_list_from_json(
featurized, model_dir=MODEL_DIR)
# Validate output files
self._check_result_file_validity()