# Main
# ---------------------------------
# Read data
data = pd.read_csv(path,
low_memory=False,
parse_dates=['date'],
usecols=features)
# Format
data = data.convert_dtypes()
data = data[features]
# Add dengue interpretation
data['dengue_interpretation'] = \
oucru_dengue_interpretation_feature(data,
pcr=True, ns1=True, igm=True,
paired_igm_igg=True, default=False)
# Overall outcome for patient
patients = data.groupby('study_no').max()
patients['month'] = patients.date.dt.month
patients['year'] = patients.date.dt.year
# Show
print("\nPatients:")
print(patients)
# Compute prevalence
aux = patients.reset_index() \
.groupby([pd.Grouper(key='date', freq='M'), 'dsource']) \
.agg(prevalence=('dengue_interpretation', prevalence),
low_memory=False,
#nrows=50000,
usecols=usecols,
parse_dates=['date'])
# Reset index
data = data.reset_index()
# Remove columns with all NaN
data = data.dropna(how='all', axis=1)
data = data.dropna(how='any', subset=features)
# Add dengue interpretation
data['dengue_interpretation'] = \
oucru_dengue_interpretation_feature(data,
pcr=True, ns1=False, igm=False, serology=False,
single_igm_igg=False, paired_igm_igg=False,
default=False)
# Create outcome
data['outcome'] = \
data.dengue_interpretation.astype(int)
# Add month
data['month'] = data.date.dt.month
# -----------
# Format data
# -----------
# Keep only day_from_admission == 0 data
#data = data[data.day_from_admission.isin([0])]
