def run_model(meta_params_dict, adjustment):
meta_params = MetaParameters()
meta_params.adjust(meta_params_dict)
policy_mods = convert_adj(adjustment["Policy"], meta_params.year.tolist())
adjustment["Tax Information"]["year"] = meta_params.year
params = CruncherParams()
params.adjust(adjustment["Tax Information"], raise_errors=False)
newvals = params.specification()
crunch = Cruncher(inputs=newvals, custom_reform=policy_mods)
#make dataset for bokeh plots
ivar = crunch.ivar
df = pd.concat([ivar] * 5000, ignore_index=True)
increments = pd.DataFrame(list(range(0, 500000, 100)))
zeros = pd.DataFrame([0] * 5000)
#ivar position of e00200p
df[9] = increments
#set spouse earning to zero
df[10] = zeros
b = Batch(df)
df_base = b.create_table()
df_reform = b.create_table(policy_mods)
#compute average tax rates
df_base['IATR'] = df_base['Individual Income Tax'] / df_base['AGI']
df_base['PATR'] = df_base['Payroll Tax'] / df_base['AGI']
df_reform['IATR'] = df_reform['Individual Income Tax'] / df_reform['AGI']
df_reform['PATR'] = df_reform['Payroll Tax'] / df_reform['AGI']
return comp_output(crunch, df_base, df_reform)
def get_inputs(meta_params_dict):
"""
Return default parameters from Tax-Cruncher
"""
metaparams = MetaParameters()
metaparams.adjust(meta_params_dict)
params = CruncherParams()
policy_params = TCParams()
keep = [
"mstat", "page", "sage", "depx", "dep13", "dep17", "dep18", "pwages",
"swages", "dividends", "intrec", "stcg", "ltcg", "otherprop",
"nonprop", "pensions", "gssi", "ui", "proptax", "otheritem",
"childcare", "mortgage", "mtr_options"
]
full_dict = params.specification(meta_data=True,
include_empty=True,
serializable=True)
params_dict = {var: full_dict[var] for var in keep}
cruncher_params = params_dict
pol_params = policy_params.specification(meta_data=True,
include_empty=True,
serializable=True,
year=metaparams.year)
meta = metaparams.specification(meta_data=True,
include_empty=True,
serializable=True)
return meta, {"Tax Information": cruncher_params, "Policy": pol_params}
def get_inputs(meta_params_dict):
"""
Return default parameters from Tax-Cruncher
"""
metaparams = MetaParameters()
metaparams.adjust(meta_params_dict)
params = CruncherParams()
policy_params = TCParams()
policy_params.set_state(year=metaparams.year.tolist())
filtered_pol_params = OrderedDict()
for k, v in policy_params.dump().items():
if k == "schema" or v.get("section_1", False):
filtered_pol_params[k] = v
keep = [
"mstat", "page", "sage", "depx", "dep13", "dep17", "dep18", "pwages",
"swages", "dividends", "intrec", "stcg", "ltcg", "otherprop",
"nonprop", "pensions", "gssi", "ui", "proptax", "otheritem",
"childcare", "mortgage", "mtr_options", "schema"
]
cruncher_dict = params.dump()
default_params = {
"Tax Information":
{k: v
for k, v in cruncher_dict.items() if k in keep},
"Policy": filtered_pol_params
}
meta = metaparams.dump()
return {"meta_parameters": meta, "model_parameters": default_params}
def run_model(meta_params_dict, adjustment):
params = CruncherParams()
params.adjust(adjustment["Tax Information"], raise_errors=False)
newvals = params.specification()
crunch = Cruncher(inputs=newvals,
custom_reform=convert_adj(adjustment["Policy"], 2019))
return comp_output(crunch)
def validate_inputs(meta_params_dict, adjustment, errors_warnings):
params = CruncherParams()
params.adjust(adjustment["Tax Information"], raise_errors=False)
errors_warnings["Tax Information"]["errors"].update(params.errors)
policy_adj = inputs.convert_policy_adjustment(adjustment["Policy"])
policy_params = Policy()
policy_params.adjust(policy_adj, raise_errors=False, ignore_warnings=True)
errors_warnings["Policy"]["errors"].update(policy_params.errors)
return {"errors_warnings": errors_warnings}
def validate_inputs(meta_params_dict, adjustment, errors_warnings):
params = CruncherParams()
params.adjust(adjustment["Tax Information"], raise_errors=False)
errors_warnings["Tax Information"]["errors"].update(params.errors)
pol_params = {}
# drop checkbox parameters.
for param, data in list(adjustment["Policy"].items()):
if not param.endswith("checkbox"):
pol_params[param] = data
policy_params = TCParams()
policy_params.adjust(pol_params, raise_errors=False)
errors_warnings["Policy"]["errors"].update(policy_params.errors)
return errors_warnings
def test_compare_defaults():
'''
Test that defaults.json and defaults_batch.json have the same
parameters in the same order.
'''
cp = CruncherParams()
ignore = ['mtr_options', 'reform_options']
cp_list = []
for k, v in cp.dump().items():
cp_list.append(k)
for item in ignore:
cp_list.remove(item)
bp = BatchParams()
bp_list = []
for k, v in bp.dump().items():
bp_list.append(k)
assert cp_list == bp_list
def get_inputs(meta_params_dict):
"""
Return default parameters from Tax-Cruncher
"""
metaparams = MetaParameters()
metaparams.adjust(meta_params_dict)
params = CruncherParams()
policy_params = Policy()
policy_params.set_state(year=metaparams.year.tolist())
policy_params.array_first = False
# Hack to work smoothly with convert_policy_defaults since
# it expects a data_source attribute.
metaparams.data_source = "CPS"
filtered_pol_params = inputs.convert_policy_defaults(
metaparams, policy_params)
keep = [
"mstat", "page", "sage", "dep13", "dep17", "dep18", "otherdep",
"pwages", "swages", "dividends", "intrec", "stcg", "ltcg", "otherprop",
"nonprop", "pensions", "gssi", "ui", "proptax", "otheritem",
"childcare", "mortgage", "businc", "sstb", "w2paid", "qualprop",
"mtr_options", "schema"
]
cruncher_dict = params.dump()
default_params = {
"Tax Information":
{k: v
for k, v in cruncher_dict.items() if k in keep},
"Policy": filtered_pol_params
}
meta = metaparams.dump()
return {"meta_parameters": meta, "model_parameters": default_params}
def run_model(meta_params_dict, adjustment):
meta_params = MetaParameters()
meta_params.adjust(meta_params_dict)
policy_mods = convert_adj(adjustment["Policy"], meta_params.year.tolist())
adjustment["Tax Information"]["year"] = meta_params.year
params = CruncherParams()
params.adjust(adjustment["Tax Information"], raise_errors=False)
newvals = params.specification()
crunch = Cruncher(inputs=newvals, custom_reform=policy_mods)
# make dataset for bokeh plots
ivar = crunch.ivar
_, mtr_opt, _ = crunch.taxsim_inputs()
df = pd.concat([ivar] * 5000, ignore_index=True)
increments = pd.DataFrame(list(range(0, 500000, 100)))
# use Calculation Option to determine what var to increment
if mtr_opt == 'Taxpayer Earnings':
span = int(ivar[9])
df[9] = increments
elif mtr_opt == 'Spouse Earnings':
span = int(ivar[10])
df[10] = increments
elif mtr_opt == 'Short Term Gains':
span = int(ivar[13])
df[13] = increments
elif mtr_opt == 'Long Term Gains':
span = int(ivar[14])
df[14] = increments
elif mtr_opt == 'Qualified Dividends':
span = int(ivar[14])
df[11] = increments
elif mtr_opt == 'Interest Received':
span = int(ivar[12])
df[12] = increments
elif mtr_opt == 'Pensions':
span = int(ivar[17])
df[17] = increments
elif mtr_opt == 'Gross Social Security Benefits':
span = int(ivar[18])
df[18] = increments
elif mtr_opt == 'Real Estate Taxes Paid':
span = int(ivar[20])
df[20] = increments
elif mtr_opt == 'Mortgage':
span = int(ivar[23])
df[23] = increments
b = Batch(df)
df_base = b.create_table()
df_reform = b.create_table(policy_mods)
# compute average tax rates
df_base['IATR'] = df_base['Individual Income Tax'] / df_base['AGI']
df_base['PATR'] = df_base['Payroll Tax'] / df_base['AGI']
df_reform['IATR'] = df_reform['Individual Income Tax'] / df_reform['AGI']
df_reform['PATR'] = df_reform['Payroll Tax'] / df_reform['AGI']
df_base['Axis'] = increments
df_reform['Axis'] = increments
return comp_output(crunch, df_base, df_reform, span, mtr_opt)