def memory_profile(s3_objs=None):
"""
Function for memory profiling pipelines with s3_objs.
Args:
s3_objs ([str]): list of s3_objs in the kitware d3batt
publication s3 bucket.
"""
s3_objs = s3_objs or MEMORY_PROFILE_S3_OBJS
# Cache s3 objects
cache_s3_objs(s3_objs, filter_existing_files=True)
with ScratchDir("."):
# Copy all pre-defined s3 objects
for obj_name in s3_objs:
shutil.copy(os.path.join(S3_CACHE, obj_name), ".")
# Snip data path prefix
data_paths = [
os.path.basename(obj) for obj in s3_objs if "Metadata" not in obj
]
# Validation
# validator = ValidatorBeep()
# validator.validate_from_paths(data_paths)
# Data structuring
raw_cycler_runs = [
RawCyclerRun.from_file(data_path) for data_path in data_paths
]
processed_cycler_runs = [
ProcessedCyclerRun.from_raw_cycler_run(raw_cycler_run)
for raw_cycler_run in raw_cycler_runs
]
# Featurization
predictors = [
DegradationPredictor.init_full_model(processed_cycler_run)
for processed_cycler_run in processed_cycler_runs
]
# Prediction
[
DegradationModel.init_full_model().predict(predictor)
for predictor in predictors
]
def test_diagnostic_feature_generation(self):
os.environ['BEEP_ROOT'] = TEST_FILE_DIR
maccor_file_w_parameters = os.path.join(
TEST_FILE_DIR, "PredictionDiagnostics_000136_00002D.037")
raw_run = RawCyclerRun.from_file(maccor_file_w_parameters)
v_range, resolution, nominal_capacity, full_fast_charge, diagnostic_available = \
raw_run.determine_structuring_parameters()
pcycler_run = ProcessedCyclerRun.from_raw_cycler_run(
raw_run, diagnostic_available=diagnostic_available)
predictor = DegradationPredictor.init_full_model(
pcycler_run,
predict_only=False,
mid_pred_cycle=11,
final_pred_cycle=12,
diagnostic_features=True)
diagnostic_feature_label = predictor.feature_labels[-1]
self.assertEqual(diagnostic_feature_label,
"median_diagnostic_cycles_discharge_capacity")
np.testing.assert_almost_equal(
predictor.X[diagnostic_feature_label][0], 4.2327690032, decimal=8)