def get_trials_group_by_folder_name():
import copy
if (get_trials_group_by_folder_name.done):
return copy.deepcopy(
get_trials_group_by_folder_name.trials_group_by_folder_name)
import load_csv_data
trials_group_by_folder_name = load_csv_data.run(
success_path=training_config.success_path,
interested_data_fields=training_config.interested_data_fields,
preprocessing_normalize=False,
preprocessing_scaling=False)
get_trials_group_by_folder_name.done = True
get_trials_group_by_folder_name.trials_group_by_folder_name = trials_group_by_folder_name
return copy.deepcopy(
get_trials_group_by_folder_name.trials_group_by_folder_name)
def get_trials_group_by_folder_name(training_config, data_class='success'):
import load_csv_data
import copy
if data_class == 'success':
data_path = training_config.success_path
elif data_class == 'test_success':
data_path = training_config.test_success_data_path
else:
raise Exception("unknown data class %s" % data_class)
trials_group_by_folder_name = load_csv_data.run(
data_path=data_path,
interested_data_fields=training_config.interested_data_fields,
preprocessing_normalize=training_config.preprocessing_normalize,
preprocessing_scaling=training_config.preprocessing_scaling,
norm_style=training_config.norm_style)
trials_group_by_folder_name
return trials_group_by_folder_name
sys.path.append("/home/birl_wu/TICC")
import TICC_solver as TICC
import numpy as np
import sys
import load_csv_data
import ipdb
base_path = 'anomaly_data_mix'
interested_data_fields = [
'.wrench_stamped.wrench.force.x',
'.wrench_stamped.wrench.force.y',
'.wrench_stamped.wrench.force.z',
'.wrench_stamped.wrench.torque.x',
'.wrench_stamped.wrench.torque.y',
'.wrench_stamped.wrench.torque.z',
]
all_trial_data = load_csv_data.run(base_path, interested_data_fields)
all_trial_data = all_trial_data.values()
for i in range(len(all_trial_data)):
if i == 0:
_temp = all_trial_data[i]
else:
_temp = np.concatenate((_temp, all_trial_data[i]), axis=0)
np.savetxt('anomaly_data.txt', _temp, delimiter=',')
fname = 'anomaly_data.txt'
(cluster_assignment, cluster_MRFs) = TICC.solve(window_size=5,
number_of_clusters=2,
lambda_parameter=11e-2,
beta=100,
maxIters=100,
threshold=2e-5,