from torch import nn
from copy import deepcopy
from src import definitions
from src.bin import statistics
from src.data_manager import cross_val
from src.models.multitask_learning import multitask_autoencoder
from src.utils.read_utils import read_pickle
feature_list = data_manager.FEATURE_LIST
# ##### Pickle #####
data_file_path = os.path.join(
definitions.DATA_DIR, 'training_data/shuffled_splits',
'training_date_normalized_shuffled_splits_select_features_no_prev_stress_2.pkl'
)
data = read_pickle(data_file_path)
############ Stats #############
print(statistics.get_train_test_val_label_counts_from_raw_data(data))
################################## Init ##################################
use_historgram = True
autoencoder_bottle_neck_feature_size = 128
autoencoder_num_layers = 1
alpha, beta = 0.001, 1
decay = 0.0001
first_key = next(iter(data['data'].keys()))
if use_historgram:
num_features = len(data['data'][first_key][4][0])
else:
num_features = len(data['data'][first_key][0][0])
def run_grid_search():
data_file_path = os.path.join(definitions.DATA_DIR,
'training_data/shuffled_splits',
TRAINING_DATA_FILE_NAME)
data = read_utils.read_pickle(data_file_path)
search_best_params_for_experiment("multitask_learner_auto_encoder", data)
eps = float(method.split('_')[1])
min_samples = int(method.split('_')[2])
feature = -1 # stress label
groups = time_warping(student_list, data, feature, eps, min_samples)
else:
groups = one_for_each(student_list)
# write to pkl file
filepath = 'Data/student_groups/' + method + '.pkl'
print('write to the file: ' + filepath)
write_utils.data_structure_to_pickle(groups, filepath)
if __name__ == '__main__':
# ##### Pickle #####
data_file_path = 'Data/training_data/shuffled_splits/training_date_normalized_shuffled_splits_select_features_no_prev_stress_all_students.pkl'
data = read_pickle(data_file_path)
student_list = conversions.extract_distinct_student_idsfrom_keys(data['data'].keys())
student_list = conversions.prepend_ids_with_string(student_list, "student_")
method = None
try:
method = sys.argv[1]
except:
method = 'one_for_each'
student_groups = clustering(student_list, data['data'], method)
groups_file_path = 'Data/student_groups/' + method + '.pkl'
print('get group file from: ' + groups_file_path)
student_groups = read_pickle(groups_file_path)
# check how students are distributed
# Start date: EST Feb.22th.2020
# Last update: EST Feb.27th.2020
# ----------------------------------------------------------------
import os
import sys
from src.utils.read_utils import read_pickle
if __name__ == '__main__':
method = None
try:
method = sys.argv[1]
except:
method = 'one_for_each'
#student_groups = clustering(student_list, data['data'], method)
groups_file_path = 'Data/student_groups/' + method + '.pkl'
print('get group file from: ' + groups_file_path)
student_groups = read_pickle(groups_file_path)
# check how students are distributed
print("student distribution: ")
rev_groups = dict()
for student in student_groups:
if rev_groups.get(student_groups[student]) != None:
rev_groups[student_groups[student]].append(student)
else:
rev_groups[student_groups[student]] = [student]
for group in rev_groups:
print(group + ': ' + str(rev_groups[group]))