taxcalc 4.2.1__py3-none-any.whl → 4.2.2__py3-none-any.whl

taxcalc/validation/taxsim35/process_taxcalc_output.py

@@ -0,0 +1,140 @@
+"""
+Translates tc --dump output file into file formatted like TAXSIM-35 output.
+"""
+
+# CODING-STYLE CHECKS:
+# pycodestyle process_tc_output.py
+# pylint --disable=locally-disabled process_tc_output.py
+
+import argparse
+import os
+import sys
+import pandas as pd
+
+
+def main(input_file_name, output_file_name):
+    """
+    Translates tc --dump output file into an output file that is
+    formatted like the first 28 variables in TAXSIM-35 output. The INPUT
+    file contains the output generated by running tc with the --dump
+    option. Any pre-existing OUTPUT file contents will be overwritten.
+    For details on Internet TAXSIM version 35 OUTPUT format, go to
+    https://users.nber.org/~taxsim/taxsim35/
+
+    Args:
+        input_file_name (string): name of file with taxcalc formatted output
+        output_file_name (string): name of file with taxsim formatted output
+
+    Returns:
+        None
+    """
+    # check INPUT filename
+    if input_file_name == "":
+        sys.stderr.write("ERROR: must specify INPUT file name\n")
+        sys.stderr.write("USAGE: {}\n".format(usage_str))
+        assert False
+    if not os.path.isfile(input_file_name):
+        emsg = "INPUT file named {} does not exist".format(args.INPUT)
+        sys.stderr.write("ERROR: {}\n".format(emsg))
+        assert False
+    # check OUTPUT filename
+    if output_file_name == "":
+        sys.stderr.write("ERROR: must specify OUTPUT file name\n")
+        sys.stderr.write("USAGE: {}\n".format(usage_str))
+        assert False
+    if os.path.isfile(output_file_name):
+        os.remove(output_file_name)
+    # read INPUT file into a pandas DataFrame
+    tcvar = pd.read_csv(input_file_name)
+    # write OUTPUT file using the pandas DataFrame
+    write_taxsim_formatted_output(output_file_name, tcvar)
+    # return no-error exit code
+    return 0
+
+
+# end of main function code
+
+
+def write_taxsim_formatted_output(filename, tcvar):
+    """
+    Write contents of tcvar pandas DataFrame to filename using
+    Internet-TAXSIM 9.3 output format containing 28 variables.
+    """
+    assert isinstance(tcvar, pd.DataFrame)
+    # with open(filename, 'w') as output_file:
+    #     for idx in range(0, len(tcvar.index)):
+    #         odict4idx = extract_output(tcvar.xs(idx))
+    #         outline = construct_output_line(odict4idx)
+    #         output_file.write(outline)
+    tcvar["state"] = 0  # state code is always zero
+    tcvar["statetax"] = 0.0  # no state income tax calculation
+    tcvar["mtr_state"] = 0.0  # no state income tax calculation
+    tcvar["zero_bracket_amount"] = (
+        0.0  # always set zero-bracket amount to zero
+    )
+    pre_phase_out_pe = tcvar["pre_c04600"].values
+    post_phase_out_pe = tcvar["c04600"].values
+    phased_out_pe = pre_phase_out_pe - post_phase_out_pe
+    tcvar["post_phase_out_pe"] = (
+        post_phase_out_pe  # post-phase-out personal exemption
+    )
+    tcvar["phased_out_pe"] = (
+        phased_out_pe  # personal exemption that is phased out
+    )
+    tcvar["exemption_surtax"] = 0.0  # always set exemption surtax to zero
+    tcvar["gen_tax_credit"] = 0.0  # always set general tax credit to zero
+    tcvar["non_refundable_child_odep_credit"] = (
+        tcvar["c07220"] + tcvar["odc"] + tcvar["ctc_new"]
+    )  # non-refundable child+odep credit
+    tcvar["refundable_CDCC"] = tcvar["CDCC_refund"]  # refundable CDCC
+    tcvar["amt_liability"] = tcvar["c09600"]  # federal AMT liability
+    # var28 from TAXSIM-35 is federal income tax before credits; the Tax-Calculator
+    # tcvar['c05800'] is this concept but includes AMT liability
+    # while Internet-TAXSIM tcvar[28] explicitly excludes AMT liability, so
+    # we have the following:
+    tcvar["iitax_before_credits_ex_AMT"] = (
+        tcvar["c05800"] - tcvar["amt_liability"]
+    )
+    tcvar = tcvar[
+        [
+            "RECID",
+            "FLPDYR",
+            "state",
+            "iitax",
+            "statetax",
+            "payrolltax",
+            "mtr_inctax",
+            "mtr_state",
+            # 'mtr_paytax',
+            "c00100",
+            "e02300",
+            "c02500",
+            # 'zero_bracket_amount',
+            "post_phase_out_pe",
+            "phased_out_pe",
+            "c21040",
+            "c04470",
+            "c04800",
+            "taxbc",
+            "exemption_surtax",
+            "gen_tax_credit",
+            "non_refundable_child_odep_credit",
+            "c11070",
+            "c07180",
+            "refundable_CDCC",
+            "eitc",
+            "c62100",
+            "amt_liability",
+            "iitax_before_credits_ex_AMT",
+            "recovery_rebate_credit",
+        ]
+    ]
+    # better mapping of to how TAXSIM-35 handles refundable credits in 2021
+    tcvar.loc[tcvar["FLPDYR"] == 2021, "c11070"] = tcvar.loc[
+        tcvar["FLPDYR"] == 2021, "non_refundable_child_odep_credit"
+    ]
+    tcvar.loc[tcvar["FLPDYR"] == 2021, "c07180"] = tcvar.loc[
+        tcvar["FLPDYR"] == 2021, "refundable_CDCC"
+    ]
+    tcvar.round(decimals=2)
+    tcvar.to_csv(filename)

taxcalc/validation/taxsim35/taxsim_emulation.json

@@ -0,0 +1,49 @@
+// JSON "reform" file that specifies changes in current-law policy that
+// are required to make Tax-Calculator work like TAXSIM-35.
+//
+// (1) AMT_child_em_c_age = 24 (rather than 18)
+//     Whether to set this parameter to 18 or 24 is arbitary because
+//     neither model has enough information to apply correctly the child
+//     AMT exemption rules.  Information on full-time student status and
+//     whether taxpayers provide more than half of their support are required
+//     to apply the rules correctly.  Tax-Calculator makes the arbitrary
+//     assumption that only those under 18 are required to use the child
+//     AMT exemption rules, while TAXSIM-27 makes the arbitrary assumption
+//     that all those under 24 are required to use the child AMT exemption.
+//     (This change was introduced for assumption set b and higher.)
+//
+// (2) EITC_excess_InvestIncome_rt = 1.0 (rather than 9e99)
+//     The rate at which the EITC amount is reduced per dollar of investment
+//     income in excess of the EITC investment income ceiling is infinity under
+//     current law (that is, any investment income in excess of the ceiling
+//     causes EITC ineligibility).  However, TAXSIM-27 assumes it is one, so
+//     that the EITC amount is reduced a dollar for each dollar of excess
+//     investment income.  This difference in the parameter value leads to
+//     many EITC differences in the randomly-generated validation samples,
+//     with some of the differences being in the thousands of dollars.  This
+//     non-current-law assumption in TAXSIM-27 is presumably made to reduce
+//     the magnitude of model-calculated marginal tax rates with respect to
+//     investment income in cases where a marginal increase in investment
+//     income takes a filing unit above the ceiling.
+
+//  (3) ALD_AlimonyReceived_hc = 1.0 (rather than 0)
+//  TAXSIM35 nonproperty income is mapped into AlimonyRecieved
+//  which had its haircut change from 1.0 to 0.0, a change that
+//  TAXSIM35 has not implemented for good reason given this:
+//      IRS: 'Beginning Jan. 1, 2019, alimony or separate
+//      maintenance payments are not deductible from the
+//      income of the payer spouse, or includable in the
+//      income of the receiving spouse, if made under a
+//      divorce or separation agreement executed after
+//      Dec. 31, 2018.
+
+// (4) PT_qbid_limit_switch = false implies TAXSIM35-like QBI deduction logic.
+{
+    "AMT_child_em_c_age": {"2013": 24},
+
+    "EITC_excess_InvestIncome_rt": {"2013": 1.0},
+
+    "ALD_AlimonyReceived_hc": {"2019": 1.0},
+
+    "PT_qbid_limit_switch": {"2018": false}
+}

taxcalc/validation/taxsim35/taxsim_input.py

@@ -0,0 +1,321 @@
+"""
+Generates random sample of tax filing units with attributes such that
+generated file can be directly uploaded to Internet TAXSIM version 35.
+"""
+
+# CODING-STYLE CHECKS:
+# pycodestyle taxsim_input.py
+# pylint --disable=locally-disabled taxsim_input.py
+
+import argparse
+import sys
+import numpy as np
+import pandas as pd
+
+
+VALID_LETTERS = ["a", "b", "c"]
+
+
+def generate_datasets(letter, year, offset=0):
+    """
+    Generates random sample of tax filing units with attributes and
+    format such that the file can be directly uploaded to Internet
+    TAXSIM version 35. For details on Internet TAXSIM version 35 INPUT
+    format, go to https://users.nber.org/~taxsim/taxsim35/
+
+    Args:
+        letter (character): letter denoting assumption set to generate data
+        year (int): year data will represent
+        offset (int): offset to alter the random number seed
+
+    Returns:
+        None
+    """
+    # check year value
+    year += 2000
+    if year < 2013 or year > 2023:
+        sys.stderr.write("ERROR: YEAR not in [2013,2023] range\n")
+        assert False
+    # check LETTER value
+    if letter == "":
+        sys.stderr.write("ERROR: must specify LETTER\n")
+        assert False
+    if letter not in VALID_LETTERS:
+        sys.stderr.write("ERROR: LETTER not in VALID_LETTERS, where\n")
+        sys.stderr.write("       VALID_LETTERS={}\n".format(VALID_LETTERS))
+        assert False
+    # check OFFSET value
+    if offset < 0 or offset > 999:
+        sys.stderr.write("ERROR: OFFSET not in [0,999] range\n")
+        assert False
+    # get dictionary containing assumption set
+    assump = assumption_set(letter, year)
+    # generate sample as pandas DataFrame
+    sample = sample_dataframe(assump, year, offset)
+    # write sample to input file
+    header_col = [
+        "taxsimid",
+        "year",
+        "state",
+        "mstat",
+        "page",
+        "sage",
+        "depx",
+        "dep13",
+        "dep17",
+        "dep18",
+        "pwages",
+        "swages",
+        "psemp",
+        "ssemp",
+        "dividends",
+        "intrec",
+        "stcg",
+        "ltcg",
+        "otherprop",
+        "nonprop",
+        "pensions",
+        "gssi",
+        "pui",
+        "sui",
+        "transfers",
+        "rentpaid",
+        "proptax",
+        "otheritem",
+        "childcare",
+        "mortgage",
+        "scorp",
+        "pbusinc",
+        "pprofinc",
+        "sbusinc",
+        "sprofinc",
+        "idtl",
+    ]
+    filename = "{}{}.in".format(letter, year % 100)
+    sample.to_csv(filename, sep=",", header=header_col, index=False)
+
+
+def assumption_set(letter, year):
+    """
+    Return dictionary containing assumption parameters.
+
+    Args:
+        letter (character): letter denoting assumption set to generate data
+        year (int): year data will represent
+
+    Returns:
+        adict (dict): assumption set dictionary (defines sampling for
+            variables)
+    """
+    adict = dict()
+    if letter in VALID_LETTERS:  # <===========================================
+        # basic assumption parameters for all ?YY.in samples:
+        adict["sample_size"] = 1000  # 100000
+        adict["year"] = year  # TAXSIM ivar 2
+        # demographic attributes:
+        adict["joint_frac"] = 0.60  # fraction of sample with joint MARS
+        adict["min_age"] = 17  # TAXSIM ivar 5 (primary taxpayer age)
+        adict["max_age"] = 77  # TAXSIM ivar 5 (primary taxpayer age)
+        adict["min_age_diff"] = -10  # min spouse age difference
+        adict["max_age_diff"] = 10  # max spouse age difference
+        adict["max_depx"] = 5  # TAXSIM ivar 7 (total number of dependents)
+        adict["max_dep13"] = 4  # TAXSIM ivar 8
+        adict["max_dep17"] = 4  # TAXSIM ivar 9
+        adict["max_dep18"] = 4  # TAXSIM ivar 10
+        # labor income:
+        adict["max_pwages_yng"] = 500  # TAXSIM ivar 11
+        adict["max_pwages_old"] = 30  # TAXSIM ivar 11 (65+ ==> old)
+        adict["max_swages_yng"] = 500  # TAXSIM ivar 12
+        adict["max_swages_old"] = 30  # TAXSIM ivar 12 (65+ ==> old)
+        # non-labor income (all zeros):
+        adict["max_psemp"] = 0  # TAXSIM ivar 13
+        adict["max_ssemp"] = 0  # TAXSIM ivar 14
+        adict["max_divinc"] = 0  # TAXSIM ivar 15
+        adict["max_intinc"] = 0  # TAXSIM ivar 16
+        adict["min_stcg"] = 0  # TAXSIM ivar 17
+        adict["max_stcg"] = 0  # TAXSIM ivar 17
+        adict["min_ltcg"] = 0  # TAXSIM ivar 18
+        adict["max_ltcg"] = 0  # TAXSIM ivar 18
+        adict["max_other_prop_inc"] = 0  # TAXSIM ivar 19
+        adict["max_other_nonprop_inc"] = 0  # TAXSIM ivar 20
+        adict["max_pnben"] = 0  # TAXSIM ivar 21
+        adict["max_ssben"] = 0  # TAXSIM ivar 22
+        adict["max_puiben"] = 0  # TAXSIM ivar 23
+        adict["max_suiben"] = 0  # TAXSIM ivar 24
+        # childcare expense amount (all zero):
+        adict["max_ccexp"] = 0  # TAXSIM ivar 29
+        # itemized expense amounts (all zero):
+        adict["max_ided_proptax"] = 0  # TAXSIM ivar 27
+        adict["max_ided_nopref"] = 0  # TAXSIM ivar 28
+        adict["max_ided_mortgage"] = 0  # TAXSIM ivar 30
+        adict["max_scorp_inc"] = 0  # TAXSIM ivar 31
+        adict["max_pbus_inc"] = 0  # TAXSIM ivar 32
+        adict["max_pprof_inc"] = 0  # TAXSIM ivar 33
+        adict["max_sbus_inc"] = 0  # TAXSIM ivar 34
+        adict["max_sprof_inc"] = 0  # TAXSIM ivar 35
+        # end if letter in VALID_LETTERS
+    if letter in ["b", "c"]:  # <==============================================
+        # non-labor income:
+        adict["max_psemp"] = 350  # TAXSIM ivar 13
+        adict["max_ssemp"] = 350  # TAXSIM ivar 14
+        adict["max_divinc"] = 20  # TAXSIM ivar 15
+        adict["max_intinc"] = 20  # TAXSIM ivar 16
+        adict["min_stcg"] = -10  # TAXSIM ivar 17
+        adict["max_stcg"] = 10  # TAXSIM ivar 17
+        adict["min_ltcg"] = -10  # TAXSIM ivar 18
+        adict["max_ltcg"] = 10  # TAXSIM ivar 18
+        adict["max_other_prop_inc"] = 30  # TAXSIM ivar 19
+        adict["max_other_nonprop_inc"] = 30  # TAXSIM ivar 20
+        adict["max_pnben"] = 60  # TAXSIM ivar 21
+        adict["max_ssben"] = 60  # TAXSIM ivar 22
+        adict["max_puiben"] = 10  # TAXSIM ivar 23
+        adict["max_suiben"] = 10  # TAXSIM ivar 24
+        adict["max_scorp_inc"] = 350  # TAXSIM ivar 31
+        adict["max_pbus_inc"] = 350  # TAXSIM ivar 32
+        adict["max_pprof_inc"] = 0  # 1  # TAXSIM ivar 33
+        adict["max_sbus_inc"] = 350  # TAXSIM ivar 34
+        adict["max_sprof_inc"] = 0  # 1  # TAXSIM ivar 35
+    if letter == "c":  # <=====================================================
+        # childcare expense amount:
+        adict["max_ccexp"] = 10  # TAXSIM ivar 29
+        # itemized expense amounts:
+        adict["max_ided_proptax"] = 30  # TAXSIM ivar 27
+        adict["max_ided_nopref"] = 10  # TAXSIM ivar 28
+        adict["max_ided_mortgage"] = 40  # TAXSIM ivar 30
+    return adict
+
+
+def sample_dataframe(assump, year, offset):
+    """
+    Construct DataFrame containing sample specified by assump and year+offset.
+
+    Args:
+        assump (dict): assumption set dictionary (defined sampling for variables)
+        year (int): year data will represent
+        offset (int): offset to alter the random number seed
+
+    Returns:
+        smpl (Pandas DataFrame): Random data in TAXSIM format
+    """
+    # pylint: disable=too-many-locals
+    np.random.seed(123456789 + year + offset)
+    size = assump["sample_size"]
+    zero = np.zeros(size, dtype=np.int64)
+    sdict = dict()
+    # (01) RECID
+    sdict[1] = range(1, size + 1)
+    # (02) YEAR
+    sdict[2] = np.full_like(zero, assump["year"], dtype=np.int64)
+    # (03) STATE
+    sdict[3] = zero
+    # (04) MSTAT
+    urn = np.random.random(size)
+    mstat = np.where(urn < assump["joint_frac"], 2, 1)
+    sdict[4] = mstat
+    # (05) PAGE
+    sdict[5] = np.random.randint(
+        assump["min_age"], assump["max_age"] + 1, size
+    )
+    # (06) SAGE
+    age_diff = np.random.randint(
+        assump["min_age_diff"], assump["max_age_diff"] + 1, size
+    )
+    sage = sdict[5] + age_diff
+    sdict[6] = np.where(mstat == 2, np.maximum(sage, assump["min_age"]), zero)
+    # (07-10) DEPX, DEP13, DEP17, DEP18
+    depx = np.random.randint(0, assump["max_depx"] + 1, size)
+    d18 = np.random.randint(0, assump["max_dep18"] + 1, size)
+    dep18 = np.where(d18 <= depx, d18, depx)
+    d17 = np.random.randint(0, assump["max_dep17"] + 1, size)
+    dep17 = np.where(d17 <= dep18, d17, dep18)
+    d13 = np.random.randint(0, assump["max_dep13"] + 1, size)
+    dep13 = np.where(d13 <= dep17, d13, dep17)
+    sdict[7] = depx
+    # (8)-(10) are ages of 3 youngest dependents
+    # If these are zero, then use depx for number of EIC children
+    # but TAXIM-35 also accepts dep13-dep18 here to be backward compatible
+    # we use that since closer to what's in tax-calculator
+    sdict[8] = dep13
+    sdict[9] = dep17
+    sdict[10] = dep18
+    # (11) PWAGES
+    pwages_yng = np.random.randint(0, assump["max_pwages_yng"] + 1, size)
+    pwages_old = np.random.randint(0, assump["max_pwages_old"] + 1, size)
+    sdict[11] = np.where(sdict[5] >= 65, pwages_old, pwages_yng) * 1000
+    # (12) SWAGES
+    swages_yng = np.random.randint(0, assump["max_swages_yng"] + 1, size)
+    swages_old = np.random.randint(0, assump["max_swages_old"] + 1, size)
+    swages = np.where(sdict[6] >= 65, swages_old, swages_yng) * 1000
+    sdict[12] = np.where(mstat == 2, swages, zero)
+    # (13) Primary Filer Self-Employment Income
+    sdict[13] = np.random.randint(0, assump["max_psemp"] + 1, size)
+    # (14) Secondary Filer Self-Employment Income
+    ssemp = np.random.randint(0, assump["max_psemp"] + 1, size)
+    sdict[14] = np.where(mstat == 2, ssemp, zero)
+    # (15) DIVIDENDS
+    sdict[15] = np.random.randint(0, assump["max_divinc"] + 1, size) * 1000
+    # (16) INTREC
+    sdict[16] = np.random.randint(0, assump["max_intinc"] + 1, size) * 1000
+    # (17) STCG
+    sdict[17] = (
+        np.random.randint(assump["min_stcg"], assump["max_stcg"] + 1, size)
+        * 1000
+    )
+    # (18) LTCG
+    sdict[18] = (
+        np.random.randint(assump["min_ltcg"], assump["max_ltcg"] + 1, size)
+        * 1000
+    )
+    # (19) OTHERPROP
+    sdict[19] = (
+        np.random.randint(0, assump["max_other_prop_inc"] + 1, size) * 1000
+    )
+    # (20) NONPROP
+    sdict[20] = (
+        np.random.randint(0, assump["max_other_nonprop_inc"] + 1, size) * 1000
+    )
+    # (21) PENSIONS
+    sdict[21] = np.random.randint(0, assump["max_pnben"] + 1, size) * 1000
+    # (22) GSSI
+    sdict[22] = np.random.randint(0, assump["max_ssben"] + 1, size) * 1000
+    # (23) Primary Filer UI (note splitting UI between primary and
+    # secondary only matters for 2020 and 2021)
+    sdict[23] = np.random.randint(0, assump["max_puiben"] + 1, size) * 1000
+    # (24) Secondary Filer UI
+    sdict[24] = np.random.randint(0, assump["max_suiben"] + 1, size) * 1000
+    # (25) TRANSFERS (non-taxable in federal income tax)
+    sdict[25] = zero
+    # (26) RENTPAID (used only in some state income tax laws)
+    sdict[26] = zero
+    # (27) PROPTAX
+    sdict[27] = (
+        np.random.randint(0, assump["max_ided_proptax"] + 1, size) * 1000
+    )
+    # (28) OTHERITEM
+    sdict[28] = (
+        np.random.randint(0, assump["max_ided_nopref"] + 1, size) * 1000
+    )
+    # (29) CHILDCARE (TAXSIM-35 EXPECTS ZERO IF NO QUALIFYING CHILDRED)
+    ccexp = np.random.randint(0, assump["max_ccexp"] + 1, size) * 1000
+    sdict[29] = np.where(dep13 > 0, ccexp, zero)
+    # (30) MORTGAGE
+    sdict[30] = (
+        np.random.randint(0, assump["max_ided_mortgage"] + 1, size) * 1000
+    )
+    # (31) S-Corp income, QBI
+    sdict[31] = np.random.randint(0, assump["max_scorp_inc"] + 1, size) * 1000
+    # (32) Primary Taxpayer's QBI
+    sdict[32] = np.random.randint(0, assump["max_pbus_inc"] + 1, size) * 1000
+    # (33) Primary Taxpayer's SSTB
+    sdict[33] = np.random.randint(0, assump["max_pprof_inc"] + 1, size)
+    # (34) Spouse's QBI
+    sqbi = np.random.randint(0, assump["max_sbus_inc"] + 1, size) * 1000
+    sdict[34] = np.where(mstat == 2, sqbi, zero)
+    # (35) Spouse's SSTB
+    spouse_sstb = np.random.randint(0, assump["max_sprof_inc"] + 1, size)
+    sdict[35] = np.where(mstat == 2, spouse_sstb, zero)
+    # (36) IDTL: variable to request intermediate calculations
+    sdict[36] = 2
+
+    smpl = pd.DataFrame(sdict)
+    return smpl

taxcalc/validation/taxsim35/tc_sims.py

@@ -0,0 +1,98 @@
+import os
+import sys
+import shutil
+import prepare_taxcalc_input as ptci
+import process_taxcalc_output as ptco
+
+CURR_PATH = os.path.abspath(os.path.dirname(__file__))
+
+
+# prepare Tax-Calculator input file
+def prep_tc_input(letter, year):
+    """
+    Args:
+        letter (character): letter denoting assumption set to generate data
+        year (int): year data will represent
+
+    Returns:
+        None
+    """
+    taxsim_in = str(letter + str(year) + ".in")
+    taxsim_in_csv = taxsim_in + ".csv"
+    ptci.main(taxsim_in, taxsim_in_csv)
+
+
+# calculate Tax-Calculator output
+def calc_tc_output(letter, year):
+    """
+    Args:
+        letter (character): letter denoting assumption set to generate data
+        year (int): year data will represent
+
+    Returns:
+        None
+    """
+    YY = str(year)
+    taxsim_in = str(letter + str(year) + ".in")
+    taxsim_in_csv = taxsim_in + ".csv"
+    taxsim_out_csv = taxsim_in + ".out.csv"
+    if os.path.exists(os.path.join(CURR_PATH, taxsim_in_csv)) is False:
+        sys.exit("ERROR: LYY_FILENAME is not a valid path")
+    year = "20" + YY
+    command = (
+        "tc "
+        + taxsim_in_csv
+        + " "
+        + year
+        + " --reform taxsim_emulation.json --dump"
+    )
+    os.system(command)
+
+    file_temp = taxsim_in + "-" + YY + "-#-taxsim_emulation-#.csv"
+    file_temp_path = os.path.join(CURR_PATH, file_temp)
+    file_out_path = os.path.join(CURR_PATH, taxsim_out_csv)
+    shutil.move(file_temp_path, file_out_path)
+
+    file_temp2 = taxsim_in + "-" + YY + "-#-taxsim_emulation-#-doc.text"
+    os.remove(file_temp2)
+
+
+def convert_to_taxsim(letter, year, save=False):
+    """
+     Convert Tax-Calculator output to TAXSIM-35 format
+
+     Args:
+        letter (character): letter denoting assumption set to generate data
+        year (int): year data will represent
+
+    Returns:
+        None
+    """
+    taxsim_in = str(letter + str(year) + ".in")
+    taxsim_in_csv = taxsim_in + ".csv"
+    taxsim_out_csv = taxsim_in + ".out.csv"
+    file_out = taxsim_in + ".out-taxcalc"
+    ptco.main(taxsim_out_csv, file_out)
+    if not save:  # Delete intermediate input and output files if not saving
+        os.remove(taxsim_in_csv)
+        os.remove(taxsim_out_csv)
+
+
+def io(letter, year):
+    """
+    Call Tax-Calculator tc CLI reading input data from specified
+    TAXSIM-35 input file and writing output in TAXSIM-35 output format
+    to a file with the specified input file name plus the .out-taxcalc
+    extension.
+
+    Args:
+        letter (character): letter denoting assumption set to generate data
+        year (int): year data will represent
+        offset (int): offset to alter the random number seed
+
+    Returns:
+        None
+    """
+    prep_tc_input(letter, year)
+    calc_tc_output(letter, year)
+    convert_to_taxsim(letter, year)

taxcalc/validation/taxsim35/tests_35.py

@@ -0,0 +1,80 @@
+import os
+import shutil
+import glob
+import input_setup
+import main_comparison
+
+CUR_PATH = os.path.abspath(os.path.dirname(__file__))
+
+# define the scope of the tests
+assumption_set = ["a", "b", "c"]  # datafiles to test
+years = [17, 18, 19, 20, 21]  # years to test
+
+# setup input files
+# if not glob.glob(os.path.join(CUR_PATH, '*in.out-taxsim')):
+input_setup.taxsim_io(assumption_set, years)
+
+# run taxcalc/taxsim comparison
+tests_passed_dict = {"a": {}, "b": {}, "c": {}}
+for letter in assumption_set:
+    for year in years:
+        tests_passed = main_comparison.main(letter, year)
+        tests_passed_dict[letter][year] = tests_passed
+
+# clean up files
+for file in CUR_PATH:
+    for file in glob.glob("*.out*") and glob.glob("*.in*"):
+        if file.endswith("taxcalc"):
+            os.remove(file)
+        if file.endswith("taxsim"):
+            os.remove(file)
+    for file in glob.glob("*.in"):
+        os.remove(file)
+# If tests passed, clean up the actual_differences directory
+# keep if tests fail to help diagnose the problem
+any_fail = False
+for letter in assumption_set:
+    for year in years:
+        if tests_passed_dict[letter][year]:
+            print(
+                "************************************************** \n"
+                + "************************************************** \n"
+                + "Validation tests for "
+                + letter
+                + str(year)
+                + " pass.  "
+                + "Any differences betweeen "
+                + "taxcalc and TAXSIM-35 are expected due to modeling "
+                + "differences. \n"
+                + "************************************************** \n"
+                + "**************************************************"
+            )
+            file = os.path.join(
+                CUR_PATH,
+                "actual_differences",
+                letter + str(year) + "differences.xlsx",
+            )
+            os.remove(file)
+            file = os.path.join(
+                CUR_PATH,
+                "actual_differences",
+                letter + str(year) + "-taxdiffs-actual.csv",
+            )
+            os.remove(file)
+
+        else:
+            any_fail = True
+            print(
+                "************************************************** \n"
+                + "************************************************** \n"
+                + "At least one validation test for "
+                + letter
+                + str(year)
+                + " failed.  Please look "
+                + "at differences in the actual and expected files and "
+                + "resolve the unexpected differences. \n"
+                + "************************************************** \n"
+                + "**************************************************"
+            )
+if not any_fail:  # if none fail, remove the actual_differences directory
+    shutil.rmtree(os.path.join(CUR_PATH, "actual_differences"))