-
Notifications
You must be signed in to change notification settings - Fork 2
/
get_label_counts.py
108 lines (84 loc) · 3.16 KB
/
get_label_counts.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
import os
import csv
import femr.labelers
import collections
import argparse
parser = argparse.ArgumentParser(
description='Get statistics about the labeling functions',
)
parser.add_argument('--path_to_data', type=str, required=True)
parser.add_argument('--path_to_output', type=str, required=True)
args = parser.parse_args()
tasks = ("PE", "1_month_mortality", "6_month_mortality", "12_month_mortality", "1_month_readmission", "6_month_readmission", "12_month_readmission", "12_month_PH")
def get_category(task):
if 'PE' in task:
return 'PE'
elif 'mortality' in task:
return 'Mort'
elif 'readmission' in task:
return 'Re-ad'
else:
return 'PH'
def get_subtype(task):
if '1_month' in task:
return '1 m'
elif '6_month' in task:
return '6 m'
elif '12_month' in task:
return '12 m'
else:
return 'N/A'
pid_split_assignment = {}
with open(os.path.join(args.path_to_data, 'inspect_cohort.csv')) as f:
reader = csv.DictReader(f)
for row in reader:
pid_split_assignment[int(row['patient_id'])] = row['split']
raw_totals = {}
for category in ('PE', 'Mort', 'Re-ad', 'PH'):
cat_tasks = [task for task in tasks if get_category(task) == category]
rows = []
for task_i, task in enumerate(cat_tasks):
labels = femr.labelers.load_labeled_patients(f'{args.path_to_output}/labels_and_features/{task}/labeled_patients.csv')
totals = collections.defaultdict(int)
positives = collections.defaultdict(int)
for pid, ls in labels.items():
split = pid_split_assignment[pid]
for l in ls:
totals[split] += 1
positives[split] += l.value
if category == 'PE':
raw_totals = totals
subrows = []
for value in ('pos.', 'neg.', 'cen.'):
row = ''
row += f'& {value}'
if value == 'pos.':
counts = positives
elif value == 'neg.':
counts = {k: totals[k] - positives[k] for k in positives}
else:
counts = {k: raw_totals[k] - totals[k] for k in positives}
if sum(counts.values()) == 0:
continue
row += f' & {sum(counts.values()):,}'
for split in ('train', 'valid', 'test'):
row += f'& {counts[split]:,} & ({ 100 * counts[split] / raw_totals[split]:0.1f} \%)'
subrows.append(row + ' \\\\')
for i, row in enumerate(subrows):
if i == 0:
row = ' & ' + '\\multirow{' + str(len(subrows)) + '}{*}{ ' + get_subtype(task) + '}' + row
else:
row = ' & ' + row
if i == len(subrows) - 1 and task_i != len(cat_tasks) - 1:
row = row + ' \n \\cline{2-10}'
subrows[i] = row
rows.extend(subrows)
for i, row in enumerate(rows):
if i == 0:
row = '\\multirow{' + str(len(rows)) + '}{*}{\\textbf{' + category + '}}' + row
else:
row = row
if i == len(rows) - 1 and task != 'PH':
row = row + ' \n \\midrule'
rows[i] = row
print('\n'.join(rows))