def extractNonLinearDomain(self, x):
try:
nni = self.extractRR(x)
sampEntro = nn.sample_entropy(nni=nni, dim=1)
lyapEx = self.lyapunov_exponent(nni=nni, emb_dim=3, matrix_dim=2)
return np.array([sampEntro["sampen"], lyapEx["lyapex"]])
except:
return np.array([])
def extractNonLinearDomain(self, x):
try:
nni = self.extractRR(x)
sampEntro = nn.sample_entropy(nni=nni, dim=2) #change dim from 1 to 2
lyapEx = self.lyapunov_exponent(nni=nni, emb_dim=3, matrix_dim=2)
pointCare = nn.poincare(nni=nni, show=False)
# hrust = nolds.hurst_rs(nni)
corrDim = nolds.corr_dim(nni, emb_dim=3)
# csi = get_csi_cvi_features(nni) #caused worse results
# geo = get_geometrical_features(nni) #caused worse results
# dfa = nolds.dfa(nni)
# return np.array([sampEntro["sampen"], lyapEx["lyapex"]])
return np.array([sampEntro["sampen"], lyapEx["lyapex"], pointCare["sd1"], pointCare["sd2"], pointCare["sd_ratio"], pointCare["ellipse_area"], corrDim])
except:
return np.array([])
def compute_features(nni):
features = {}
features['mean_hr'] = tools.heart_rate(nni).mean()
features['sdnn'] = td.sdnn(nni)[0]
features['rmssd'] = td.rmssd(nni)[0]
features['sdsd'] = td.sdsd(nni)[0]
features['nn20'] = td.nn20(nni)[0]
features['pnn20'] = td.nn20(nni)[1]
features['nn50'] = td.nn50(nni)[0]
features['pnn50'] = td.nn50(nni)[1]
features['hf_lf_ratio'] = fd.welch_psd(nni, show=False)['fft_ratio']
features['very_lf'] = fd.welch_psd(nni, show=False)['fft_peak'][0]
features['lf'] = fd.welch_psd(nni, show=False)['fft_peak'][1]
features['hf'] = fd.welch_psd(nni, show=False)['fft_peak'][2]
features['log_very_lf'] = fd.welch_psd(nni, show=False)['fft_log'][0]
features['log_lf'] = fd.welch_psd(nni, show=False)['fft_log'][1]
features['log_hf'] = fd.welch_psd(nni, show=False)['fft_log'][2]
features['sampen'] = nl.sample_entropy(nni)[0]
return features
def cal_hrv(data_list):
if (np.nan in data_list) or (len(data_list) <= 1):
sdsd = np.nan
rmssd = np.nan
sd1 = np.nan
sd2 = np.nan
sd_ratio = np.nan
a1 = np.nan
a2 = np.nan
a_ratio = np.nan
sampen = np.nan
else:
# sdsd
sdsd = round(td.sdsd(nni=data_list)['sdsd'], 5)
# rmssd
rmssd = round(td.rmssd(nni=data_list)['rmssd'], 5)
# sd1, sd2
nl_results = nl.poincare(nni=data_list)
sd1 = round(nl_results['sd1'], 5)
sd2 = round(nl_results['sd2'], 5)
sd_ratio = round(sd2 / sd1, 5)
# dfa a1, a2
dfa_results = nl.dfa(data_list)
try:
a1 = round(dfa_results['dfa_alpha1'], 5)
a2 = round(dfa_results['dfa_alpha2'], 5)
a_ratio = round(a2 / a1, 5)
except:
a1 = np.nan
print(a1, type(a1))
a2 = np.nan
print(a2, type(a2))
a_ratio = np.nan
# Sampen
t = np.std(data_list)
sampen = round(
nl.sample_entropy(nni=data_list, tolerance=t)['sampen'], 6)
return sdsd, rmssd, sd1, sd2, sd_ratio, a1, a2, a_ratio, sampen
def updateInfo():
# sql server setting and connect
IPadr = 'DESKTOP-01CEBMI' # Server ip address
srcDatabase = '' # DB name
Database_ID = '' # SQL ID
Database_PWD = '' # SQL PW
q = pyodbc.connect(DRIVER='{SQL Server}',
SERVER=IPadr,
DATABASE=srcDatabase,
UID=Database_ID,
PWD=Database_PWD)
Query = """ select * from matching_patient
order by delirium desc, CONVERT(int, subclass) asc
"""
df = pd.io.sql.read_sql(Query, q)
Query = """ select * from results_latest
order by delirium desc, CONVERT(int, subclass) asc
"""
df_smoking = pd.io.sql.read_sql(Query, q)
# numeric variable
# numeric_vars = ['VITAL_HR', 'VITAL_SPO2', 'NIBP_DIAS', 'NIBP_SYS', 'VITAL_RESP']
parameters = ['ECGII', 'RESPIMP', 'SPO2IRAC']
onset_time = []
delirium_id = []
control_id = []
for d in range(0, len(df['delirium_onset'])
): # devide delirium patient ID and Control patient ID
if df['delirium_onset'][d] == '0': # delirium_onset=0 : control
control_id.append(df['patientid'][d])
else: # else : delirium
delirium_id.append(df['patientid'][d])
for time_series in time_series_list:
for parameter in parameters:
if os.path.isfile(dstDir + "/" + parameter + "_data_" +
str(time_series // 60) + "H.json") != 1:
open(
dstDir + "/" + parameter + "_data_" +
str(time_series // 60) + "H.json", 'a').close()
data = dict()
data['pid'] = []
data['subclass'] = []
data['delirium'] = []
data['onset_time'] = []
data['ICU_in'] = []
data['LOS'] = []
data['starttime'] = []
data['endtime'] = []
data['smoking'] = []
# sample entropy
data['sampen'] = []
# DFA
data['DFA_a1'] = []
data['DFA_a2'] = []
data['DFA_ratio'] = []
else:
f = open(
dstDir + "/" + parameter + "_data_" +
str(time_series // 60) + "H.json", "r")
data = f.read()
data = json.loads(data)
f.close()
idx = -1
c_data = 0 # Control cohort data exist patient's number
d_data = 0 # Delirium(Case) cohort data exist patient's number
c_no_data = 0 # Control cohort no data patient's number
d_no_data = 0 # Delirium(Case) cohort no data patient's number
for i in df['patientid']:
idx += 1
p_id = i
sub_class = df['subclass'][idx]
smoking = int(df_smoking['smoking(ex0)']
[idx]) # 0: non-smoking, 1: smoking
data['smoking'].append(smoking)
if p_id in data['pid']:
new_id = p_id + "_2"
data['pid'].append(new_id)
else:
data['pid'].append(p_id)
data['subclass'].append(int(sub_class))
# extract data each patients
data_list = []
if list(df['delirium_onset']
)[idx] == '0': # control group patients
# create empty array
nni_result = np.array([])
ICU_in = list(df['Time_in'])[idx]
LOS = list(df['LOS'])[idx]
ICU_time = datetime.datetime.strptime(ICU_in, '%Y%m%d%H%M')
data['LOS'].append(float(LOS))
data['ICU_in'].append(int(ICU_in))
case_LOS = float(list(df['LOS'])[int(sub_class) - 1])
case_onset = float(onset_time[int(sub_class) - 1])
print('case_LOS : ', case_LOS)
print('case_onset : ', case_onset)
onset_ratio = round(case_onset / case_LOS, 3)
print('onset_ratio : ', onset_ratio)
control_LOS = float(LOS)
control_onset = round(control_LOS * onset_ratio, 3)
control_onset_time = round(control_onset * 24 * 60, 0)
print('control_LOS : ', control_LOS)
print('control_onset : ', control_onset)
print('')
# onset_control = round(onset_time[int(sub_class)-1] * 24 * 60,0) # 24hour, 60minutes
data['onset_time'].append(control_onset)
t1 = (ICU_time +
timedelta(minutes=(control_onset_time - time_series))
).strftime('%Y%m%d%H%M') # ICU_in_time + onset_time
t1 = t1[:11] + '0'
t2 = datetime.datetime.strptime(t1, '%Y%m%d%H%M')
t3 = (t2 - timedelta(minutes=(measure_time))
).strftime('%Y%m%d%H%M')
print('')
print('control onset time: ', t1)
print('onset - 30min : ', t3)
print('')
data['starttime'].append(int(t3))
data['endtime'].append(int(t1))
# select data ---> time 기준 이전 30분 데이터를 뽑아서 내림차순 정렬 후, 가장 최근 데이터를 추출 (직전 10분 데이터가 없는 경우를 대비)
Query = """
select top 1 filepath
from waveform_data
where patient_id= '""" + p_id + """'
and left(starttime, 12) >= '""" + str(
t3) + """'
and left(starttime, 12) < '""" + str(
t1) + """'
and wavetype = '""" + parameter + """'
order by id desc
"""
df1 = pd.io.sql.read_sql(Query, q)['filepath']
if df1.shape[0] == 0: # if data not exist,
c_no_data += 1
print('')
print('no control data', 'pid: ', p_id)
print('')
if df['delirium'][idx] == '0':
data['delirium'].append(0)
else:
data['delirium'].append(1)
data['sampen'].append(0)
data['DFA_a1'].append(0)
data['DFA_a2'].append(0)
data['DFA_ratio'].append(0) # alpha 1 / alpha 2
else:
if parameter == 'ECGII':
if df['delirium'][idx] == '0':
data['delirium'].append(0)
else:
data['delirium'].append(1)
c_data += 1
file_path = "E://backup/waveform_data/BiosignalRepository/BiosignalRepository/GZIP/" + str(
df1[0][28:]).replace("\\", "/")
patient_id = str(df1[0][28:48])
date = str(df1[0][49:57])
f_name = str(df1[0][58:])
print('')
print('file_path: ', file_path)
print('')
data_10m = read_nk(file_path, parameter)[:]
try:
if ((data_10m == 500).all()) or ((data_10m
== 0).all()):
print('data not Found', 'pid: ', p_id)
file_path = "E://backup/waveform_data/BiosignalRepository/BiosignalRepository/GZIP/" + str(
df1[1][28:]).replace("\\", "/")
data_10m = read_nk(file_path, parameter)[:]
else:
data_10m = read_nk(file_path, parameter)[:]
except:
data_10 = read_nk(file_path, parameter)[:]
t = np.std(data_10m) # tolerance = t * 0.2
sampen_result = nl.sample_entropy(data_10m,
dim=2,
tolerance=t * 0.2)
dfa_result = nl.dfa(data_10m, show=False) # dim 바꿔봐야함
sampen = round(sampen_result[0], 5)
dfa_a1 = round(dfa_result[1], 5) # [1] is alpha1
dfa_a2 = round(dfa_result[2], 5) # [2] is alpha2
dfa_ratio = round(dfa_a1 / dfa_a2,
5) # ratio is a1 / a2
data['sampen'].append(sampen)
data['DFA_a1'].append(dfa_a1)
data['DFA_a2'].append(dfa_a2)
data['DFA_ratio'].append(dfa_ratio)
print('sampen: ', sampen)
print('')
print('a1: ', dfa_a1)
print('a2: ', dfa_a2)
print('ratio: ', dfa_ratio)
else: # case group patients
# create empty array
ICU_in = list(df['Time_in'])[idx]
ICU_time = datetime.datetime.strptime(
ICU_in[0:11] + '0', '%Y%m%d%H%M')
LOS = list(df['LOS'])[idx]
data['ICU_in'].append(ICU_in)
data['LOS'].append(float(LOS))
onset = list(
df['delirium_onset']
)[idx] # onset > if delirium patient exist, find onset_time index
# print(onset)
delirium_onset = float(df['onset_time'][idx])
data['onset_time'].append(delirium_onset)
onset_time.append(delirium_onset)
onset_date = datetime.datetime.strptime(
onset[0:11] + '0', '%Y%m%d%H%M') # delirium_onset_time
print('onset_date:', onset_date)
t4 = (onset_date -
timedelta(minutes=(time_series))).strftime(
'%Y%m%d%H%M') # str (onset time)
t5 = datetime.datetime.strptime(
t4, '%Y%m%d%H%M'
) # onset time # datetime (modified onset time)
t6 = (t5 - timedelta(minutes=(measure_time))).strftime(
'%Y%m%d%H%M') # str (onset_time - measure time)
# t5 = (onset_date).strftime('%Y%m%d%H%M')
print('')
print('delirium onset time: ', t4)
print('onset - 30min : ', t6)
print('')
data['starttime'].append(int(t6))
data['endtime'].append(int(t4))
# select data ---> time 기준 이전 30분 데이터를 뽑아서 내림차순 정렬 후, 가장 최근 데이터를 추출 (직전 10분 데이터가 없는 경우를 대비)
Query = """
select top 1 filepath
from waveform_data
where patient_id= '""" + p_id + """'
and left(starttime, 12) >= '""" + str(
t6) + """'
and left(starttime, 12) < '""" + str(
t4) + """'
and wavetype = '""" + parameter + """'
order by id desc
"""
df2 = pd.io.sql.read_sql(Query, q)['filepath']
# ob_value2 = list(map(float, df2[:]))
# ob_value2 = [i for i in ob_value2 if i != 0]
if df2.shape[
0] == 0: # if exist less than 10% for measure period time series,
d_no_data += 1
print('')
print('no case data', 'pid: ', p_id)
print('')
if df['delirium'][idx] == '0':
data['delirium'].append(0)
else:
data['delirium'].append(1)
data['sampen'].append(0)
data['DFA_a1'].append(0)
data['DFA_a2'].append(0)
data['DFA_ratio'].append(0) # alpha 1 / alpha 2
else:
if parameter == 'ECGII':
if df['delirium'][idx] == '0':
data['delirium'].append(0)
else:
data['delirium'].append(1)
d_data += 1
file_path = "E://backup/waveform_data/BiosignalRepository/BiosignalRepository/GZIP/" + str(
df2[0][28:]).replace("\\", "/")
patient_id = str(df2[0][28:48])
date = str(df2[0][49:57])
f_name = str(df2[0][58:])
print('')
print('file_path: ', file_path)
print('')
data_10m = read_nk(file_path, parameter)[:]
try:
if ((data_10m == 500).all()) or ((data_10m
== 0).all()):
print('data not Found', 'pid: ', p_id)
file_path = "E://backup/waveform_data/BiosignalRepository/BiosignalRepository/GZIP/" + str(
df1[1][28:]).replace("\\", "/")
data_10m = read_nk(file_path, parameter)[:]
else:
data_10m = read_nk(file_path, parameter)[:]
except:
data_10 = read_nk(file_path, parameter)[:]
t = np.std(data_10m) # tolerance = t * 0.2
sampen_result = nl.sample_entropy(data_10m,
dim=2,
tolerance=t * 0.2)
dfa_result = nl.dfa(data_10m, show=False) # dim 바꿔봐야함
sampen = round(sampen_result[0], 5)
dfa_a1 = round(dfa_result[1], 5) # [1] is alpha1
dfa_a2 = round(dfa_result[2], 5) # [2] is alpha2
dfa_ratio = round(dfa_a1 / dfa_a2,
5) # ratio is a1 / a2
data['sampen'].append(sampen)
data['DFA_a1'].append(dfa_a1)
data['DFA_a2'].append(dfa_a2)
data['DFA_ratio'].append(dfa_ratio)
print('sampen: ', sampen)
print('a1: ', dfa_a1)
print('a2: ', dfa_a2)
print('ratio: ', dfa_ratio)
print('')
print('delirium_data: ', d_data)
print('control_data: ', c_data)
print('delirium_data loss: ', d_no_data)
print('control_data loss: ', c_no_data)
print('origin_patient: ', len(df['delirium_onset']))
print('total data: ', c_data + d_data + c_no_data + d_no_data)
print('')
print('')
# data_df = pd.DataFrame(data)
# data_df.to_csv (dstDir + "/spo2_data_" + str(measure_time // 60) + ".csv", index = False, header=True)
with open(dstDir + "/" + parameter + "_data_" +
str(time_series // 60) + "H.json",
'w',
encoding='utf-8') as f:
json.dump(data, f, indent="\t")
f.close()