def prepare_GSC (control_file, num_days_list):
print ('start prepare_GSC')
for num_days in num_days_list:
handle_signal_data_new.create_lab_data(control_file,
lab_file=data_dir + 'mimic_gsc_score.csv',
lab_parameter_name='gsc', num_days=num_days,
output_dir_in_func=output_dir)
def prepare_breathing (control_file,num_days_list):
blood_file = data_dir+ 'blood_data_for_modeling_clean.csv'
print ('start prepare_breathing')
num_days = 100 # this is used also for apache - first day in icu so we need all the data
# -------10/01/2019 - no need - data in mimic is in the same hours
#todo- delete comment
fio2_joined_file = join_fio2_sources(control_file,blood_file=blood_file,respiratory_file=data_dir + 'Respiratory_data_for_modeling.csv')
fio2_joined_file = output_dir + 'fio2_both_sources.csv'
create_fio2_pio2_ratio(control_file,
pio2_file=blood_file,
fio2_file=fio2_joined_file,
num_days=num_days,
new_param_name='FIO2',default_val=30)
handle_signal_data_new.create_A_aPO2(paco2_file=blood_file,
paco2_param_name = 'pCO2',
fio2_file = output_dir+'FIO2_calculated.csv', fio2_param_name = 'FIO2',
pao2_file=blood_file,pao2_param_name = 'pO2')
# ---------end no need
#todo - delete comment
df_to_add = handle_signal_data_new.create_signal_combination(blood_file,
'pO2', 'FIO2', 'divide', fio2_joined_file,min_val_feature2=21)
for num_days in num_days_list:
handle_signal_data_new.create_lab_data(control_file,
lab_file=output_dir + 'pO2_divide_FIO2.csv',
lab_parameter_name='pO2_divide_FIO2', num_days=num_days,
output_dir_in_func=output_dir)
handle_signal_data_new.create_lab_data(control_file,
lab_file=output_dir + 'A_aPO2.csv',
lab_parameter_name='A_aPO2', num_days=num_days, output_dir_in_func=output_dir,
complete_0_values=True, complete_hours=True)
def prepare_blood(control_file, blood_lab_file = data_dir+'blood_data_for_modeling.csv',num_days_list=[5]):
blood_lab_file_clean = data_dir+'blood_data_for_modeling_clean.csv'
blood_df = pd.read_csv(blood_lab_file, encoding="ISO-8859-1")
blood_df = blood_df[blood_df['value'].str.contains('>|<|GREATER|ERROR|UNABLE|LESS|VOIDED|-') == False]
blood_df['value'] = blood_df['value'].str.replace('+','')
blood_df = blood_df[blood_df['value'].apply(lambda x: x[0].isdigit())]
blood_df.to_csv(blood_lab_file_clean)
group_functions = {'patient_id' : {'patient_id_nunique': 'nunique', 'rows_cnt' :'count'}}
blood_df_grp = blood_df.groupby(['parameter_name']).agg(group_functions)
blood_df_grp.columns = blood_df_grp.columns.droplevel(0)
blood_df_grp.reset_index(level=0, inplace=True)
# print (blood_df_grp)
blood_df_grp_filter = blood_df_grp[blood_df_grp.patient_id_nunique>200]
#take only numeric values
for param_name in blood_df_grp_filter.parameter_name :
if param_name in ['Ventilation Rate','Anisocytosis','Fibrin Degradation Products']:
continue
# if param_name < 'Hypochromia':
# print ('skip '+param_name)
# continue
for num_days in num_days_list:
handle_signal_data_new.create_lab_data(control_file,
lab_df=blood_df,
lab_parameter_name=param_name,
num_days=num_days, output_dir_in_func=output_dir)
def prepare_signals(control_file, data_file = data_dir+'Respiratory_data_for_modeling.csv',num_days_list=[5]):
print ('start prepare_signals - '+data_file)
data_df = pd.read_csv(data_file, encoding="ISO-8859-1")
data_df = data_df.rename(columns = {'SUBJECT_ID':'patient_id','VALUE':'value','CHARTTIME':'time','LABEL':'parameter_name'})
# data_df.to_csv(data_file)
group_functions = {'patient_id' : {'patient_id_nunique': 'nunique', 'rows_cnt' :'count'}}
data_df_grp = data_df.groupby(['parameter_name']).agg(group_functions)
data_df_grp.columns = data_df_grp.columns.droplevel(0)
data_df_grp.reset_index(level=0, inplace=True)
print (data_df_grp)
blood_df_grp_filter = data_df_grp[data_df_grp.patient_id_nunique>200]
#take only numeric values
for param_name in blood_df_grp_filter.parameter_name :
calc_std_flag = True
smooth_signal = True
use_fft_feature = True
if param_name=='Heart Rate':
min_value = 35
max_value = 210
elif param_name == 'Temperature Fahrenheit':
min_value = 93.2
max_value = 113
elif param_name=='Temperature Celsius':
min_value = 34.5
max_value = 45
elif param_name in ['Arterial Blood Pressure mean','ART BP mean','Non Invasive Blood Pressure mean']:
min_value = 25
max_value = 330
elif param_name in ['Arterial Blood Pressure diastolic','ART BP Diastolic','Non Invasive Blood Pressure diastolic',
'Manual Blood Pressure Diastolic Right']:
min_value = 20
max_value = 310
elif param_name in ['ART BP Systolic','Arterial Blood Pressure systolic','Non Invasive Blood Pressure systolic',
'Manual Blood Pressure Systolic Right','Manual Blood Pressure Systolic Left']:
min_value = 30
max_value = 350
elif param_name=='Respiratory Rate':
min_value = 4
max_value = 80
else:
min_value = -99999
max_value = 99999
calc_std_flag = False
smooth_signal = False
use_fft_feature = False
for num_days in num_days_list:
try:
handle_signal_data_new.create_lab_data(control_file,
lab_df=data_df,
lab_parameter_name=param_name,min_value=min_value,
max_value=max_value,
smooth_signal=smooth_signal,use_fft_feature=use_fft_feature
,calc_std_flag=calc_std_flag,
num_days=num_days, output_dir_in_func=output_dir)
except Exception:
print(param_name + ' - create_lab_data failed')
pass
def prepare_medicine (control_file, num_days_list):
print ('start prepare_medicine')
#---vasoactive-------------
org_file = data_dir+'mimic_medications_data_for_modeling.csv'
new_file_name = output_dir+'mimic_medications_data_with_dummy.csv'
updated_vasoactive_file = handle_signal_data_new.add_dummy_rows_for_control_set(control_file,
org_file,
'Norepinephrine',
output_file=new_file_name)
updated_vasoactive_file = handle_signal_data_new.add_dummy_rows_for_control_set(control_file,
updated_vasoactive_file,
'Phenylephrine',
output_file=new_file_name)
updated_vasoactive_file = handle_signal_data_new.add_dummy_rows_for_control_set(control_file,
updated_vasoactive_file,
'Epinephrine',
output_file=new_file_name)
updated_vasoactive_file = handle_signal_data_new.add_dummy_rows_for_control_set(control_file,
updated_vasoactive_file,
'Dopamine',
output_file=new_file_name)
updated_vasoactive_file = handle_signal_data_new.add_dummy_rows_for_control_set(control_file,
updated_vasoactive_file,
'Vasopressin',
output_file=new_file_name)
for num_days in num_days_list:
handle_signal_data_new.create_lab_data(control_file,
lab_file=updated_vasoactive_file,
lab_parameter_name='Norepinephrine', num_days=num_days, complete_0_values=True, complete_hours=True, output_dir_in_func=output_dir)
handle_signal_data_new.create_lab_data(control_file,
lab_file=updated_vasoactive_file,
lab_parameter_name='Phenylephrine', num_days=num_days, complete_0_values=True, complete_hours=True, output_dir_in_func=output_dir)
handle_signal_data_new.create_lab_data(control_file,
lab_file=updated_vasoactive_file,
lab_parameter_name='Epinephrine', num_days=num_days, complete_0_values=True, complete_hours=True, output_dir_in_func=output_dir)
handle_signal_data_new.create_lab_data(control_file,
lab_file=updated_vasoactive_file,
lab_parameter_name='Dopamine', num_days=num_days, complete_0_values=True, complete_hours=True, output_dir_in_func=output_dir)
handle_signal_data_new.create_lab_data(control_file,
lab_file=updated_vasoactive_file,
lab_parameter_name='Vasopressin', num_days=num_days, complete_0_values=True, complete_hours=True, output_dir_in_func=output_dir)
#------------Morphine/Fentanyl/Propofol 2% ( Diprivan )' (only 1 patient with 1%), 'Dormicum'
updated_vasoactive_file = handle_signal_data_new.add_dummy_rows_for_control_set(control_file,
org_file,
'Midazolam(Versed)',
output_file=new_file_name)
updated_vasoactive_file = handle_signal_data_new.add_dummy_rows_for_control_set(control_file,
updated_vasoactive_file,
'Fentanyl',
output_file=new_file_name)
updated_vasoactive_file = handle_signal_data_new.add_dummy_rows_for_control_set(control_file,
updated_vasoactive_file,
'Morphine Sulfate',
output_file=new_file_name)
updated_vasoactive_file = handle_signal_data_new.add_dummy_rows_for_control_set(control_file,
updated_vasoactive_file,
'Propofol',
output_file=new_file_name)
for num_days in num_days_list:
# todo - delete comment
handle_signal_data_new.create_lab_data(control_file,
lab_file=updated_vasoactive_file,
lab_parameter_name='Fentanyl', num_days=num_days, complete_0_values=True, complete_hours=True, output_dir_in_func=output_dir)
handle_signal_data_new.create_lab_data(control_file,
lab_file=updated_vasoactive_file,
lab_parameter_name='Morphine Sulfate', num_days=num_days, complete_0_values=True, complete_hours=True, output_dir_in_func=output_dir)
handle_signal_data_new.create_lab_data(control_file,
lab_file=updated_vasoactive_file,
lab_parameter_name='Propofol', num_days=num_days, complete_0_values=True, complete_hours=True, output_dir_in_func=output_dir)
handle_signal_data_new.create_lab_data(control_file,
lab_file=updated_vasoactive_file,
lab_parameter_name='Midazolam(Versed)', num_days=num_days, complete_0_values=True, complete_hours=True, output_dir_in_func=output_dir)