forked from harp-lab/brainPH
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathdistance_calculation.py
215 lines (193 loc) · 10.1 KB
/
distance_calculation.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
import argparse
import json
import os
import gudhi
import gudhi.wasserstein
import numpy as np
from sklearn.manifold import MDS
from utils import get_dataset, timer
from barcodes_calculation import get_0_dim_barcodes
def get_mds(dissimilarity_matrix, is_euclidean=None, random_state=6):
if is_euclidean:
embedding = MDS(n_components=2, random_state=random_state)
else:
embedding = MDS(n_components=2, dissimilarity="precomputed",
random_state=random_state)
return embedding.fit_transform(dissimilarity_matrix)
def get_barcodes_distance(dgm_1, dgm_2, distance_method='ws'):
if distance_method == 'ws':
return gudhi.wasserstein.wasserstein_distance(dgm_1, dgm_2,
order=1., internal_p=2.)
elif distance_method == 'bn':
return gudhi.bottleneck_distance(dgm_1, dgm_2)
elif distance_method == 'ed':
return np.linalg.norm(dgm_1 - dgm_2)
def get_barcodes_single_subject(data_dir, subject_number, manual=False):
print(f"Calculating barcodes for Subject {subject_number}")
filepath_645 = f'{data_dir}/subject_{subject_number}_mx645.txt'
filepath_1400 = f'{data_dir}/subject_{subject_number}_mx1400.txt'
filepath_2500 = f'{data_dir}/subject_{subject_number}_std2500.txt'
normalized_matrix_645 = get_dataset(filename=filepath_645, fmri=True)
normalized_matrix_1400 = get_dataset(filename=filepath_1400, fmri=True)
normalized_matrix_2500 = get_dataset(filename=filepath_2500, fmri=True)
if manual:
barcodes_645 = np.array(get_0_dim_barcodes(normalized_matrix_645))
barcodes_1400 = np.array(get_0_dim_barcodes(normalized_matrix_1400))
barcodes_2500 = np.array(get_0_dim_barcodes(normalized_matrix_2500))
else:
rips_complex_645 = gudhi.RipsComplex(
distance_matrix=normalized_matrix_645)
rips_complex_1400 = gudhi.RipsComplex(
distance_matrix=normalized_matrix_1400)
rips_complex_2500 = gudhi.RipsComplex(
distance_matrix=normalized_matrix_2500)
pd_645 = rips_complex_645.create_simplex_tree(
max_dimension=1).persistence()[1:]
pd_1400 = rips_complex_1400.create_simplex_tree(
max_dimension=1).persistence()[1:]
pd_2500 = rips_complex_2500.create_simplex_tree(
max_dimension=1).persistence()[1:]
barcodes_645 = np.array([pair[1] for pair in pd_645])
barcodes_1400 = np.array([pair[1] for pair in pd_1400])
barcodes_2500 = np.array([pair[1] for pair in pd_2500])
return [barcodes_645, barcodes_1400, barcodes_2500]
def get_barcodes(data_directory, total_subjects):
barcodes = []
for subject_number in range(1, total_subjects + 1):
barcodes.append(get_barcodes_single_subject(data_directory,
subject_number))
return barcodes
def get_distances(barcodes, subject_number, distance_method='ws'):
barcodes_645, barcodes_1400, barcodes_2500 = barcodes[subject_number - 1]
distance_645_1400 = get_barcodes_distance(barcodes_645,
barcodes_1400,
distance_method=distance_method)
distance_1400_2500 = get_barcodes_distance(barcodes_1400,
barcodes_2500,
distance_method=distance_method)
distance_645_2500 = get_barcodes_distance(barcodes_645,
barcodes_2500,
distance_method=distance_method)
return [round(distance_645_1400, 3),
round(distance_1400_2500, 3),
round(distance_645_2500, 3)]
@timer
def compute_distances_between_cohorts(barcodes,
total_subjects,
start_subject=None,
end_subject=None,
distance_method='ws',
data_dir="full_data", output_dir="output"):
if not os.path.exists(output_dir):
os.makedirs(output_dir)
if start_subject == None:
start_subject = 1
end_subject = total_subjects
generated_json = f'{output_dir}/distances_between_cohorts_{distance_method}.json'
distances = []
for subject_number in range(start_subject, end_subject + 1):
print(f"Calculating distances between cohorts "
f"for Subject {subject_number}")
distances.append(get_distances(barcodes,
subject_number,
distance_method))
with open(generated_json, "w") as f:
json.dump(distances, f)
print(
f"Done generating the {distance_method} JSON file between cohorts: {generated_json}")
@timer
def compute_mds_within_a_cohort(barcodes, total_subjects, cohort,
distance_method='ws', generate_file=True,
output_directory='output'):
if not os.path.exists(output_directory):
os.makedirs(output_directory)
cohort_name = {0: "mx645", 1: "mx1400", 2: "std2500"}[cohort]
print(f"Calculating {distance_method} distance matrix "
f"of {total_subjects} subjects for cohort {cohort_name}")
dissimilarity_matrix = np.array([[0.0 for j in range(total_subjects)]
for i in range(total_subjects)])
for i in range(total_subjects):
barcodes_1 = barcodes[i][cohort]
for j in range(i):
barcodes_2 = barcodes[j][cohort]
distance = get_barcodes_distance(barcodes_1,
barcodes_2,
distance_method=distance_method)
distance = round(distance, 3)
dissimilarity_matrix[i - 1][j - 1] = distance
dissimilarity_matrix[j - 1][i - 1] = distance
print(f"Calculating MDS of {total_subjects} subjects "
f"for cohort {cohort_name}")
if distance_method == "ed":
mds_matrix = get_mds(dissimilarity_matrix, is_euclidean=True)
else:
mds_matrix = get_mds(dissimilarity_matrix, is_euclidean=False,
random_state=258)
if generate_file:
generated_matrix_file = f'{output_directory}/distance_matrix_{cohort_name}_{distance_method}.json'
with open(generated_matrix_file, "w") as f:
json.dump(dissimilarity_matrix.tolist(), f)
print(f"Done generating {generated_matrix_file}")
generated_mds_file = f'{output_directory}/mds_{cohort_name}_{distance_method}.json'
with open(generated_mds_file, "w") as f:
json.dump(mds_matrix.tolist(), f)
print(f"Done generating {generated_mds_file}")
@timer
def compute_mds_of_all_cohorts(barcodes, total_subjects,
distance_method='ws', data_dir="full_data", output_dir="output"):
compute_mds_within_a_cohort(barcodes, total_subjects, 0,
distance_method, output_directory=output_dir)
compute_mds_within_a_cohort(barcodes, total_subjects, 1,
distance_method, output_directory=output_dir)
compute_mds_within_a_cohort(barcodes, total_subjects, 2,
distance_method, output_directory=output_dir)
def get_user_input():
parser = argparse.ArgumentParser()
parser.add_argument('--method', '-m',
help='Enter one of the distance method (ws, bn)')
parser.add_argument('--start', '-s',
help='Enter start subject (1, 316)')
parser.add_argument('--end', '-e',
help='Enter end subject (1, 316)')
parser.add_argument('--distance', '-p',
help='To calculate distance matrix (y or n)')
parser.add_argument('--mds', '-q',
help='To calculate MDS (y or n)')
parser.add_argument('--data_dir', '-d',
help='Enter input data folder (e.g. full_data)')
parser.add_argument('--output_dir', '-o',
help='Enter output data folder (e.g. output)')
args = parser.parse_args()
if args.start:
main(args.method, start_subject=int(args.start),
end_subject=int(args.end),
distance_calculation=args.distance,
mds_calculation=args.mds,
data_dir=args.data_dir,
output_dir=args.output_dir)
return
parser.print_help()
@timer
def main(method, start_subject=1, end_subject=316,
distance_calculation='y', mds_calculation='y', data_dir="full_data", output_dir="output"):
total_subjects = 316
barcodes = get_barcodes(data_dir, total_subjects)
if distance_calculation == 'y':
compute_distances_between_cohorts(barcodes,
total_subjects,
start_subject,
end_subject,
distance_method=method,
data_dir=data_dir,
output_dir=output_dir)
if mds_calculation == 'y':
compute_mds_of_all_cohorts(barcodes, total_subjects,
distance_method=method,
data_dir=data_dir,
output_dir=output_dir)
if __name__ == "__main__":
get_user_input()
# python distance_calculation.py --method ws --start 1 --end 316 --distance y --mds y --data_dir full_data_linear --output_dir output_linear
# python distance_calculation.py --method ws --start 1 --end 316 --distance y --mds y --data_dir full_data_positive_linear --output_dir output_positive_linear
# python distance_calculation.py --method ws --start 1 --end 316 --distance y --mds y --data_dir full_data_negative_linear --output_dir output_negative_linear
# python distance_calculation.py --method ws --start 1 --end 316 --distance y --mds y --data_dir full_data_linear_downsample --output_dir output_linear_downsample