def save_clusters(y, offset, filename):
i = 0
y = y + offset
center_list = datasets.load_centers_all()
for center in center_list:
for person in center.persons:
if person.label == 2:
person.label = y[i]
i += 1
for center in center_list:
center.save_labels(filename)
#%%
import center
import datasets
import numpy as np
import mask
import nibabel as nib
from sklearn.preprocessing import OneHotEncoder
centers = datasets.load_centers_all()
def get_index(lst, item):
return [i for i in range(len(lst)) if lst[i] > item]
def onehot(n_labels, labels):
return np.eye(n_labels)[labels]
def minmax(x, axis=0):
_min = np.min(x, axis=axis)
_max = np.max(x, axis=axis)
return (x - _min) / (_max - _min)
def z_norm(x, axis=0):
_mean = np.mean(x, axis=axis)
_std = np.std(x, axis=axis)
return (x - _mean) / _std
#%%
import datasets
center_list = datasets.load_centers_all()
# %%
#prepare HYDRA csv
import csv
import numpy as np
feature_path = './matlab/HYDRA/data/features.csv'
covariate_path = './matlab/HYDRA/data/covariate.csv'
i = 0
with open(feature_path, 'w', newline='') as feature_file:
header = ['id'] + ['feature_{}'.format(i) for i in range(456)] + ['group']
featureswriter = csv.DictWriter(feature_file, fieldnames=header)
featureswriter.writeheader()
with open(covariate_path, 'w', newline='') as covariate_file:
covariatewriter = csv.writer(covariate_file)
for center in center_list:
features_rv, labels = center.get_csv_values(flatten=True)
features_ct, _ = center.get_csv_values(
prefix='cortical_thickness/{}.csv', flatten=True)
features = np.concatenate([features_rv, features_ct], axis=1)
personal_infos, _ = center.get_presonal_info_values()
tcgws, _ = center.get_tivs_cgws()
for feature, label, personal_info, tcgw in zip(
features, labels, personal_infos, tcgws):
if label == 2:
label = 1
elif label == 0:
label = -1
else:
continue