def run_nth_year_tax_calc_model(year_n,
start_year,
use_puf_not_cps,
use_full_sample,
user_mods,
return_dict=True):
"""
The run_nth_year_tax_calc_model function assumes user_mods is a dictionary
returned by the Calculator.read_json_param_objects() function.
Setting use_puf_not_cps=True implies use puf.csv input file;
otherwise, use cps.csv input file.
Setting use_full_sample=False implies use sub-sample of input file;
otherwsie, use the complete sample.
"""
# pylint: disable=too-many-arguments,too-many-locals
start_time = time.time()
# create calc1 and calc2 calculated for year_n and mask
check_years_return_first_year(year_n, start_year, use_puf_not_cps)
(calc1, calc2, mask) = calculate(year_n,
start_year,
use_puf_not_cps,
use_full_sample,
user_mods,
behavior_allowed=True)
# extract raw results from calc1 and calc2
rawres1 = results(calc1.records)
rawres2 = results(calc2.records)
# seed random number generator with a seed value based on user_mods
seed = random_seed(user_mods)
print('seed={}'.format(seed))
np.random.seed(seed) # pylint: disable=no-member
# construct TaxBrain summary results from raw results
summ = summary(rawres1, rawres2, mask)
def append_year(pdf):
"""
append_year embedded function revises all column names in pdf
"""
pdf.columns = [str(col) + '_{}'.format(year_n) for col in pdf.columns]
return pdf
# optionally return non-JSON-like results
if not return_dict:
res = dict()
for tbl in summ:
res[tbl] = append_year(summ[tbl])
elapsed_time = time.time() - start_time
print('elapsed time for this run: {:.1f}'.format(elapsed_time))
return res
# optionally construct JSON-like results dictionaries for year n
dec_row_names_n = [x + '_' + str(year_n) for x in DECILE_ROW_NAMES]
bin_row_names_n = [x + '_' + str(year_n) for x in WEBBIN_ROW_NAMES]
agg_row_names_n = [x + '_' + str(year_n) for x in AGG_ROW_NAMES]
dist_column_types = [float] * len(DIST_TABLE_LABELS)
diff_column_types = [float] * len(DIFF_TABLE_LABELS)
info = dict()
for tbl in summ:
info[tbl] = {'row_names': [], 'col_types': []}
if 'dec' in tbl:
info[tbl]['row_names'] = dec_row_names_n
elif 'bin' in tbl:
info[tbl]['row_names'] = bin_row_names_n
else:
info[tbl]['row_names'] = agg_row_names_n
if 'dist' in tbl:
info[tbl]['col_types'] = dist_column_types
elif 'diff' in tbl:
info[tbl]['col_types'] = diff_column_types
res = dict()
for tbl in summ:
if 'aggr' in tbl:
res_table = create_dict_table(summ[tbl],
row_names=info[tbl]['row_names'])
res[tbl] = dict((k, v[0]) for k, v in res_table.items())
else:
res[tbl] = create_dict_table(summ[tbl],
row_names=info[tbl]['row_names'],
column_types=info[tbl]['col_types'])
elapsed_time = time.time() - start_time
print('elapsed time for this run: {:.1f}'.format(elapsed_time))
return res
def run_nth_year_taxcalc_model(year_n,
start_year,
use_puf_not_cps,
use_full_sample,
user_mods,
return_dict=True):
"""
The run_nth_year_taxcalc_model function assumes user_mods is a dictionary
returned by the Calculator.read_json_param_objects() function.
Setting use_puf_not_cps=True implies use puf.csv input file;
otherwise, use cps.csv input file.
Setting use_full_sample=False implies use sub-sample of input file;
otherwsie, use the complete sample.
"""
# pylint: disable=too-many-arguments,too-many-locals,too-many-branches
start_time = time.time()
# create calc1 and calc2 calculated for year_n
check_years_return_first_year(year_n, start_year, use_puf_not_cps)
calc1, calc2 = calculate(year_n,
start_year,
use_puf_not_cps,
use_full_sample,
user_mods,
behavior_allowed=True)
# extract unfuzzed raw results from calc1 and calc2
dv1 = calc1.distribution_table_dataframe()
dv2 = calc2.distribution_table_dataframe()
# delete calc1 and calc2 now that raw results have been extracted
del calc1
del calc2
# construct TaxBrain summary results from raw results
sres = dict()
fuzzing = use_puf_not_cps
if fuzzing:
# seed random number generator with a seed value based on user_mods
# (reform-specific seed is used to choose whose results are fuzzed)
seed = random_seed(user_mods)
print('fuzzing_seed={}'.format(seed))
np.random.seed(seed) # pylint: disable=no-member
# make bool array marking which filing units are affected by the reform
reform_affected = np.logical_not( # pylint: disable=no-member
np.isclose(dv1['combined'], dv2['combined'], atol=0.01, rtol=0.0))
agg1, agg2 = fuzzed(dv1, dv2, reform_affected, 'aggr')
sres = summary_aggregate(sres, agg1, agg2)
del agg1
del agg2
dv1b, dv2b = fuzzed(dv1, dv2, reform_affected, 'xbin')
sres = summary_dist_xbin(sres, dv1b, dv2b)
sres = summary_diff_xbin(sres, dv1b, dv2b)
del dv1b
del dv2b
dv1d, dv2d = fuzzed(dv1, dv2, reform_affected, 'xdec')
sres = summary_dist_xdec(sres, dv1d, dv2d)
sres = summary_diff_xdec(sres, dv1d, dv2d)
del dv1d
del dv2d
del reform_affected
else:
sres = summary_aggregate(sres, dv1, dv2)
sres = summary_dist_xbin(sres, dv1, dv2)
sres = summary_diff_xbin(sres, dv1, dv2)
sres = summary_dist_xdec(sres, dv1, dv2)
sres = summary_diff_xdec(sres, dv1, dv2)
# nested function used below
def append_year(pdf):
"""
append_year embedded function revises all column names in pdf
"""
pdf.columns = [str(col) + '_{}'.format(year_n) for col in pdf.columns]
return pdf
# optionally return non-JSON-like results
if not return_dict:
res = dict()
for tbl in sres:
res[tbl] = append_year(sres[tbl])
elapsed_time = time.time() - start_time
print('elapsed time for this run: {:.1f}'.format(elapsed_time))
return res
# optionally construct JSON-like results dictionaries for year n
dec_rownames = list(sres['diff_comb_xdec'].index.values)
dec_row_names_n = [x + '_' + str(year_n) for x in dec_rownames]
bin_rownames = list(sres['diff_comb_xbin'].index.values)
bin_row_names_n = [x + '_' + str(year_n) for x in bin_rownames]
agg_row_names_n = [x + '_' + str(year_n) for x in AGG_ROW_NAMES]
dist_column_types = [float] * len(DIST_TABLE_LABELS)
diff_column_types = [float] * len(DIFF_TABLE_LABELS)
info = dict()
for tbl in sres:
info[tbl] = {'row_names': [], 'col_types': []}
if 'dec' in tbl:
info[tbl]['row_names'] = dec_row_names_n
elif 'bin' in tbl:
info[tbl]['row_names'] = bin_row_names_n
else:
info[tbl]['row_names'] = agg_row_names_n
if 'dist' in tbl:
info[tbl]['col_types'] = dist_column_types
elif 'diff' in tbl:
info[tbl]['col_types'] = diff_column_types
res = dict()
for tbl in sres:
if 'aggr' in tbl:
res_table = create_dict_table(sres[tbl],
row_names=info[tbl]['row_names'])
res[tbl] = dict((k, v[0]) for k, v in res_table.items())
else:
res[tbl] = create_dict_table(sres[tbl],
row_names=info[tbl]['row_names'],
column_types=info[tbl]['col_types'])
elapsed_time = time.time() - start_time
print('elapsed time for this run: {:.1f}'.format(elapsed_time))
return res