def get_survival_measure_data(data):
data = pivot_data(
data[results.RESULT_COLUMNS('survival_alive') +
results.RESULT_COLUMNS('survival_other') + GROUPBY_COLUMNS])
for s in reversed(RISKS):
data['process'], data[s] = data.process.str.split('_' + s + '_').str
data['measure'], data['process'] = data.process.str.split('_period_').str
data['period'], data['treatment_line'] = data.process.str.split(
'_line_').str
data = data.drop(columns='process')
return sort_data(data)
def get_measure_data(data,
measure,
stratified_by_treatment=False,
stratified_by_risks=False):
data = pivot_data(data[results.RESULT_COLUMNS(measure) + GROUPBY_COLUMNS])
data = split_processing_column(data, stratified_by_treatment,
stratified_by_risks)
return sort_data(data)
def get_treatment_count_measure_data(data, measure):
data = pivot_data(data[results.RESULT_COLUMNS(measure) + GROUPBY_COLUMNS])
data['process'], data['year'] = data.process.str.split('_year_').str
data['process'], data['treatment'] = data.process.str.split(
'_treatment_').str
data['process'], data['treatment_line'] = data.process.str.split(
'line_').str
data = data.drop(columns='process')
return sort_data(data)
def get_registry_measure_data(data):
data = pivot_data(data[results.RESULT_COLUMNS('registry_status') +
GROUPBY_COLUMNS])
data['process'], data[
'renal_function_at_diagnosis'] = data.process.str.split(
'_renal_function_at_diagnosis_').str
data['process'], data[
'race_and_cytogenetic_risk_at_diagnosis'] = data.process.str.split(
'_race_and_cytogenetic_risk_at_diagnosis_').str
data['process'], data['age'] = data.process.str.split('_in_age_group_').str
data['process'], data['sex'] = data.process.str.split('_among_').str
data['year'] = data.process.str.split('_in_').str[-1]
data['measure'] = data.process.str.split('_in_').str[:-1].apply(
lambda x: '_in_'.join(x))
data = data.drop(columns='process')
return sort_data(data)
def aggregate_over_seed(data):
non_count_columns = []
for non_count_template in results.NON_COUNT_TEMPLATES:
non_count_columns += results.RESULT_COLUMNS(non_count_template)
count_columns = [
c for c in data.columns if c not in non_count_columns + GROUPBY_COLUMNS
]
# non_count_data = data[non_count_columns + GROUPBY_COLUMNS].groupby(GROUPBY_COLUMNS).mean()
count_data = data[count_columns +
GROUPBY_COLUMNS].groupby(GROUPBY_COLUMNS).sum()
return pd.concat(
[
count_data,
# non_count_data
],
axis=1).reset_index()
def get_population_data(data):
total_pop = pivot_data(
data[[results.TOTAL_POPULATION_COLUMN] +
results.RESULT_COLUMNS('population') + GROUPBY_COLUMNS])
total_pop = total_pop.rename(columns={'process': 'measure'})
return sort_data(total_pop)