def total_historic_nnb(dic_data, input_dic, idx=general.month_end_series):
'''
Function that returns a total historical nnb based on the implied nnb on months that we do not have the total nnb data
dic_data: fund data
input_dic: dictionary of inputs which include nnb distribution
'''
implied_discretionary_nnb = get_historic_implied_nnb(dic_data,
idx=idx) # series
monthly_filled_inputs = general.monthly_fulfill(input_dic)
total_historical_nnb = dic_data['total nnb'].reindex(index=idx)
implied_total_nnb = (
implied_discretionary_nnb /
monthly_filled_inputs['nnb distribution'].reindex(
index=idx)[general.disc_known_cols].sum(axis='columns')).to_frame(
name='NNB') # dataframe
'''
Pandas version 0.21 fixed the following bug!
result = implied_total_nnb[total_historical_nnb.isnull()].fillna(0.0) + total_historical_nnb.fillna(0.0)
'''
result = pandas.concat([
total_historical_nnb[~numpy.isnan(total_historical_nnb.values)],
implied_total_nnb[numpy.isnan(total_historical_nnb.values)]
],
axis='index')
return result
def total_nnb_distribution_clientAlgo(dic_data, input_dic, opt=None):
total = total_nnb_clientAlgo(dic_data, input_dic)
result = (general.monthly_fulfill(input_dic)['nnb distribution']).reindex(
index=general.month_end_series).multiply(total['NNB'], axis='index')
result.iloc[0, :] = 0
if opt is None:
return result
else:
if opt == 'financial_year' or opt == 'calendar_year':
return general.convert_fy_quarter_half_index(
result, result.index).groupby(opt).sum(min_count=1)
elif opt == 'month_no':
return general.convert_fy_quarter_half_index(
result, result.index).groupby(['calendar_year',
opt]).sum(min_count=1)
else:
return general.convert_fy_quarter_half_index(
result, result.index).groupby(['financial_year',
opt]).sum(min_count=1)
def future_nnb_distribution(dic_data, input_dic):
total = total_future_nnb(dic_data, input_dic)
result = (general.monthly_fulfill(input_dic)['nnb distribution']).reindex(
index=general.month_end_series).multiply(total['NNB'], axis='index')
result.iloc[0, :] = 0
return result