forked from landiisotta/behavioral_phenotyping
-
Notifications
You must be signed in to change notification settings - Fork 0
/
dataset.py
237 lines (201 loc) · 7.27 KB
/
dataset.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
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
from sqlalchemy import create_engine, MetaData
import datetime
from datetime import datetime
import csv
import os
import utils as ut
import pandas as pd
import logging
from dataclasses import dataclass
from basic_statistics import DataStatistics
import numpy as np
# Dataclasses to store patient demographics,
# and patient info on encounters.
@dataclass
class Pinfo:
sex: str
dob: str
n_enc: int = 0
@dataclass
class Penc:
sex: str
dob: str
doa_instrument: list()
def count_enc(self):
yr_enc = list(map(lambda x: x[0].split('/')[2],
self.doa_instrument))
return len(set(yr_enc))
# Configure the logging, logging to file.
logging.basicConfig(level=logging.INFO,
filename='./logs/pipeline.log',
filemode='w')
# Create new directory or point to an existing one to store data.
data_dir = 'odf-data'
data_path = os.path.join(ut.DATA_FOLDER_PATH, data_dir)
os.makedirs(data_path, exist_ok=True)
runtime_date = datetime.now().strftime('%Y-%m-%d-%H-%M-%S')
logging.info(f'{runtime_date} created ../data/odf-data folder for returned objects')
def access_db():
""" Access the database and dump tables.
Returns
-------
dictionary
{key=table_name, value=pandas dataframe}
"""
# connect to the database
engine = create_engine(ut.SQLALCHEMY_CONN_STRING)
conn = engine.connect()
logging.info('Connection to DB established')
# inspect the tables in the database
metadata = MetaData(engine, reflect=True)
logging.info('Dumping all tables')
df_tables = {}
for table_name in metadata.tables:
# ADDED THIS TO STOP IMPORTING NEW DATA FOR NOW
df_tables[table_name] = pd.read_sql_table(table_name,
con=conn,
parse_dates=['date_birth', 'date_ass'],
index_col='id').query(
'form_info <= datetime(2019, 10, 5)').drop('form_info', axis=1)
return df_tables
def data_wrangling(tables_dict):
""" Drop excluded subjects and tables
Parameters
----------
tables_dict: dictionary
dictionary with dumped tables from DB
Returns
-------
dictionary
reduced dictionary without excluded tables and subjects (rows)
"""
adult_subj = tables_dict['ados-2modulo4'].id_subj.unique()
# added lab1680 on the 1st of October 2019, new entry with only WISC-IV
# added also lab1353/lab1152, only psi-sf/srs available
adult_subj = np.append(adult_subj, ['lab1680', 'lab1353', 'lab1152'])
logging.info(f'Dropped {len(adult_subj)} subjects')
# names of the tables to drop from the dictionary
tb_drop = ['ados-2modulo4',
'emotionalavailabilityscales']
tb_dict_rid = {}
for tb_name, df in tables_dict.items():
if tb_name not in tb_drop:
row_drop = ~(df['id_subj'].isin(adult_subj))
tb_dict_rid[tb_name] = df.loc[row_drop]
return tb_dict_rid
def cohort_info(tables_dict):
"""Store instances of Pinfo and Penc classes in dictionaries
Parameters
----------
tables_dict: dictionary
dictionary with data tables
Returns
-------
dictionary
{keys=pid, values=Pinfo instances}
dictionary
{keys=pid, values=Penc instances}
"""
demog_dict = {}
enc_dict = {}
for tn, df in tables_dict.items():
for _, row in df.iterrows():
ass_date = __correct_datetime(row.date_ass)
birth_date = __correct_datetime(row.date_birth)
if row.id_subj in enc_dict:
enc_dict[row.id_subj].doa_instrument.append((ass_date, tn))
else:
enc_dict[row.id_subj] = Penc(sex=row.sex,
dob=birth_date,
doa_instrument=[(ass_date,
tn)])
demog_dict[row.id_subj] = Pinfo(sex=row.sex,
dob=birth_date)
for pid in demog_dict:
demog_dict[pid].n_enc = enc_dict[pid].count_enc()
# dump info to csv files
_dump_info(demog_dict, enc_dict)
# save log with statistics
logging.info('\nComputing basics statistics (DataStatistics module)\n')
DataStatistics().compute(data_dir)
return demog_dict, enc_dict
"""
Functions
"""
def age_ass(dob, doa):
"""
Parameters
----------
dob: str
date of birth
doa: str
date of assessment
Return
------
float
age of assessment
"""
# dob = pd.Timestamp(year=int(dob.split('/')[2]),
# month=int(dob.split('/')[1]),
# day=int(dob.split('/')[0]))
# doa = pd.Timestamp(year=int(doa.split('/')[2]),
# month=int(doa.split('/')[1]),
# day=int(doa.split('/')[0]))
dob = pd.Timestamp(dob)
doa = pd.Timestamp(doa)
days_in_year = 365.2425
aoa = (doa - dob).days / days_in_year
return aoa
def __correct_datetime(date_ts):
"""
Parameters
----------
date_ts: pandas Timestamp
Returns
-------
str
strftime %d/%m/%Y
"""
# correct wrong dates
today = datetime.today()
try:
if date_ts.year == today.year and date_ts.month >= today.month:
corrected_date = pd.Timestamp(year=date_ts.year,
month=date_ts.day,
day=date_ts.month)
else:
corrected_date = date_ts
return corrected_date.strftime("%d/%m/%Y")
except AttributeError:
return date_ts
def _dump_info(demog_info, enc_info):
"""Save csv file with demographic and encounter info
Parameters
----------
demog_info: dictionary
{keys=pid, values=Pinfo instances}
enc_info: dictionary
{keys=pid, values=Penc instances}
"""
logging.info("Saving csv files on subject info and subject encounters")
with open(os.path.join(ut.DATA_FOLDER_PATH, data_dir,
'person-encounters.csv'), 'w') as f:
wr = csv.writer(f, delimiter=',', quoting=csv.QUOTE_MINIMAL)
wr.writerow(['ID_SUBJ', 'SEX', 'DOB', 'DOA', 'AOA', 'INSTRUMENT'])
for pid in sorted(enc_info.keys()):
enc_info[pid].doa_instrument.sort(key=lambda x: (x[0].split('/')[2],
x[0].split('/')[1],
x[1]))
for tup in enc_info[pid].doa_instrument:
wr.writerow([pid, enc_info[pid].sex,
enc_info[pid].dob, tup[0],
age_ass(enc_info[pid].dob, tup[0]),
tup[1]])
with open(os.path.join(ut.DATA_FOLDER_PATH, data_dir,
'person-demographics.csv'), 'w') as f:
wr = csv.writer(f, delimiter=',', quoting=csv.QUOTE_MINIMAL)
wr.writerow(['ID_SUBJ', 'SEX', 'DOB', 'N_ENC'])
for pid in sorted(demog_info.keys()):
wr.writerow([pid, demog_info[pid].sex,
demog_info[pid].dob,
demog_info[pid].n_enc])