def read_icd_diagnoses_table(mimic3_path):
codes = dataframe_from_csv(os.path.join(mimic3_path, 'D_ICD_DIAGNOSES.csv'))
codes = codes[['ICD9_CODE', 'SHORT_TITLE', 'LONG_TITLE']]
diagnoses = dataframe_from_csv(os.path.join(mimic3_path, 'DIAGNOSES_ICD.csv'))
diagnoses = diagnoses.merge(codes, how='inner', left_on='ICD9_CODE', right_on='ICD9_CODE')
diagnoses[['SUBJECT_ID', 'HADM_ID', 'SEQ_NUM']] = diagnoses[['SUBJECT_ID', 'HADM_ID', 'SEQ_NUM']].astype(int)
return diagnoses
def read_icd_diagnoses_table(mimic3_path):
"""
concatenate 2 csv into 1 DataFrame
`D_ICD_DIAGNOSES.csv` : 'ICD9_CODE', 'SHORT_TITLE', 'LONG_TITLE'
&
`DIAGNOSES_ICD.csv` : 'SUBJECT_ID', 'HADM_ID', 'SEQ_NUM'
"""
codes = dataframe_from_csv(os.path.join(mimic3_path,
'D_ICD_DIAGNOSES.csv'))
codes = codes[['ICD9_CODE', 'SHORT_TITLE', 'LONG_TITLE']]
diagnoses = dataframe_from_csv(
os.path.join(mimic3_path, 'DIAGNOSES_ICD.csv'))
# merging 2 DF
diagnoses = diagnoses.merge(codes,
how='inner',
left_on='ICD9_CODE',
right_on='ICD9_CODE')
# dtype : int
diagnoses[['SUBJECT_ID', 'HADM_ID',
'SEQ_NUM']] = diagnoses[['SUBJECT_ID', 'HADM_ID',
'SEQ_NUM']].astype(int)
return diagnoses
def read_admissions_table(mimic3_path):
admits = dataframe_from_csv(os.path.join(mimic3_path, 'ADMISSIONS.csv'))
admits = admits[['SUBJECT_ID', 'HADM_ID', 'ADMITTIME', 'DISCHTIME', 'DEATHTIME', 'ETHNICITY', 'DIAGNOSIS']]
admits.ADMITTIME = pd.to_datetime(admits.ADMITTIME)
admits.DISCHTIME = pd.to_datetime(admits.DISCHTIME)
admits.DEATHTIME = pd.to_datetime(admits.DEATHTIME)
return admits
def read_stays(subject_path):
stays = dataframe_from_csv(os.path.join(subject_path, 'stays.csv'), index_col=None)
stays.INTIME = pd.to_datetime(stays.INTIME)
stays.OUTTIME = pd.to_datetime(stays.OUTTIME)
stays.DOB = pd.to_datetime(stays.DOB)
stays.DOD = pd.to_datetime(stays.DOD)
stays.DEATHTIME = pd.to_datetime(stays.DEATHTIME)
stays.sort_values(by=['INTIME', 'OUTTIME'], inplace=True)
return stays
def read_events(subject_path, remove_null=True):
events = dataframe_from_csv(os.path.join(subject_path, 'events.csv'), index_col=None)
if remove_null:
events = events[events.VALUE.notnull()]
events.CHARTTIME = pd.to_datetime(events.CHARTTIME)
events.HADM_ID = events.HADM_ID.fillna(value=-1).astype(int)
events.ICUSTAY_ID = events.ICUSTAY_ID.fillna(value=-1).astype(int)
events.VALUEUOM = events.VALUEUOM.fillna('').astype(str)
# events.sort_values(by=['CHARTTIME', 'ITEMID', 'ICUSTAY_ID'], inplace=True)
return events
def read_icustays_table(mimic3_path):
"""
from `mimi3_path` read `ICUSTAYS.csv` as DataFrame and only return
['SUBJECT_ID', 'HADM_ID', 'ICUSTAY_ID', 'DBSOURCE', 'FIRST_CAREUNIT','LAST_CAREUNIT', 'FIRST_WARDID', 'LAST_WARDID','INTIME', 'OUTTIME','LOS']
`date` convert to dtype `datetime`
"""
stays = dataframe_from_csv(os.path.join(mimic3_path, 'ICUSTAYS.csv'))
stays.INTIME = pd.to_datetime(stays.INTIME)
stays.OUTTIME = pd.to_datetime(stays.OUTTIME)
return stays
def read_itemid_to_variable_map(fn, variable_column='LEVEL2'):
var_map = dataframe_from_csv(fn, index_col=None).fillna('').astype(str)
# var_map[variable_column] = var_map[variable_column].apply(lambda s: s.lower())
var_map.COUNT = var_map.COUNT.astype(int)
var_map = var_map[(var_map[variable_column] != '') & (var_map.COUNT > 0)]
var_map = var_map[(var_map.STATUS == 'ready')]
var_map.ITEMID = var_map.ITEMID.astype(int)
var_map = var_map[[variable_column, 'ITEMID',
'MIMIC LABEL']].set_index('ITEMID')
return var_map.rename(
{
variable_column: 'VARIABLE',
'MIMIC LABEL': 'MIMIC_LABEL'
}, axis=1)
def read_patients_table(mimic3_path):
"""
from `mimi3_path` read `PATENTS.csv` as DataFrame and only return
'SUBJECT_ID', 'GENDER', 'DOB', 'DOD'
these 4 columns
`date` convert to dtype `datetime`
"""
pats = dataframe_from_csv(os.path.join(mimic3_path, 'PATIENTS.csv'))
pats = pats[['SUBJECT_ID', 'GENDER', 'DOB', 'DOD']]
pats.DOB = pd.to_datetime(pats.DOB)
pats.DOD = pd.to_datetime(pats.DOD)
return pats
def read_admissions_table(mimic3_path):
"""
from `mimi3_path` read `ADMISSIONS.csv` as DataFrame and only return
'SUBJECT_ID', 'HADM_ID', 'ADMITTIME', 'DISCHTIME', 'DEATHTIME', 'ETHNICITY', 'DIAGNOSIS'
these 7 columns
`date` convert to dtype `datetime`
"""
admits = dataframe_from_csv(os.path.join(mimic3_path, 'ADMISSIONS.csv'))
admits = admits[[
'SUBJECT_ID', 'HADM_ID', 'ADMITTIME', 'DISCHTIME', 'DEATHTIME',
'ETHNICITY', 'DIAGNOSIS'
]]
admits.ADMITTIME = pd.to_datetime(admits.ADMITTIME)
admits.DISCHTIME = pd.to_datetime(admits.DISCHTIME)
admits.DEATHTIME = pd.to_datetime(admits.DEATHTIME)
return admits
def read_icustays_table(mimic3_path):
stays = dataframe_from_csv(os.path.join(mimic3_path, 'ICUSTAYS.csv'))
stays.INTIME = pd.to_datetime(stays.INTIME)
stays.OUTTIME = pd.to_datetime(stays.OUTTIME)
return stays
def read_patients_table(mimic3_path):
pats = dataframe_from_csv(os.path.join(mimic3_path, 'PATIENTS.csv'))
pats = pats[['SUBJECT_ID', 'GENDER', 'DOB', 'DOD']]
pats.DOB = pd.to_datetime(pats.DOB)
pats.DOD = pd.to_datetime(pats.DOD)
return pats
def read_diagnoses(subject_path):
return dataframe_from_csv(os.path.join(subject_path, 'diagnoses.csv'),
index_col=None)
phenotypes = add_hcup_ccs_2015_groups(
diagnoses, yaml.load(open(args.phenotype_definitions, 'r')))
make_phenotype_label_matrix(phenotypes,
stays).to_csv(os.path.join(args.output_path,
'phenotype_labels.csv'),
index=False,
quoting=csv.QUOTE_NONNUMERIC)
if args.test:
pat_idx = np.random.choice(patients.shape[0], size=1000)
patients = patients.iloc[pat_idx]
stays = stays.merge(patients[['SUBJECT_ID']],
left_on='SUBJECT_ID',
right_on='SUBJECT_ID')
args.event_tables = [args.event_tables[0]]
print('Using only', stays.shape[0], 'stays and only', args.event_tables[0],
'table')
subjects = stays.SUBJECT_ID.unique()
break_up_stays_by_subject(stays, args.output_path, subjects=subjects)
break_up_diagnoses_by_subject(phenotypes, args.output_path, subjects=subjects)
items_to_keep = set([
int(itemid)
for itemid in dataframe_from_csv(args.itemids_file)['ITEMID'].unique()
]) if args.itemids_file else None
for table in args.event_tables:
read_events_table_and_break_up_by_subject(args.mimic3_path,
table,
args.output_path,
items_to_keep=items_to_keep,
subjects_to_keep=subjects)
