Derive raw-PUF partnership self-employment earnings with the archived formula

changelog.d/fix-raw-puf-partnership-se-fallback.fixed.md

@@ -0,0 +1 @@
+Derive `partnership_self_employment_net_earnings` from a raw PUF with the archived pipeline's Schedule SE residual instead of the sum of partnership and S corporation losses, and fail closed when a raw PUF lacks any of its eight input columns. The certified default reads the processed donor's `partnership_se_income` column and is unchanged.

docs/evidence/spec-engine/us-f0-coverage.json

@@ -1656,13 +1656,13 @@
           "compiler_ir.node_slices"
         ],
         "expected": {
-          "map_sha256": "bdd339868ee0cec07fa3a52a91c4bd97de707c15aeb0dc224b064be9ce04eb4b",
-          "protocol_sha256": "8d028354d487eadecf3bc4a950d70c95e5c44c792fe7fdf2d52c77e6d86a7886"
+          "map_sha256": "41a64b8820bf161e316c49e3f2b666a247b58f6f6d50dd4395b89c8f5b1399f6",
+          "protocol_sha256": "0e4c8a4ace5e10747065669a009691d0eee9179ca42b344f8912c5ca239ebde6"
         },
         "failures": [],
         "observed": {
-          "map_sha256": "bdd339868ee0cec07fa3a52a91c4bd97de707c15aeb0dc224b064be9ce04eb4b",
-          "protocol_sha256": "8d028354d487eadecf3bc4a950d70c95e5c44c792fe7fdf2d52c77e6d86a7886"
+          "map_sha256": "41a64b8820bf161e316c49e3f2b666a247b58f6f6d50dd4395b89c8f5b1399f6",
+          "protocol_sha256": "0e4c8a4ace5e10747065669a009691d0eee9179ca42b344f8912c5ca239ebde6"
         },
         "status": "covered"
       },
@@ -1677,7 +1677,7 @@
           "compiler_ir.seed_stream_map"
         ],
         "expected": {
-          "implementation_sha256": "8d028354d487eadecf3bc4a950d70c95e5c44c792fe7fdf2d52c77e6d86a7886",
+          "implementation_sha256": "0e4c8a4ace5e10747065669a009691d0eee9179ca42b344f8912c5ca239ebde6",
           "protocol": "legacy-v1",
           "streams": [
             "build_model",
@@ -1698,7 +1698,7 @@
         },
         "failures": [],
         "observed": {
-          "implementation_sha256": "8d028354d487eadecf3bc4a950d70c95e5c44c792fe7fdf2d52c77e6d86a7886",
+          "implementation_sha256": "0e4c8a4ace5e10747065669a009691d0eee9179ca42b344f8912c5ca239ebde6",
           "protocol": "legacy-v1",
           "streams": [
             "build_model",
@@ -2599,7 +2599,7 @@
       "country": "us",
       "schema_id": "country_spec",
       "schema_version": 1,
-      "spec_sha256": "638cd60e97a9b200f8bdfd4f92e67be986f516ac25b9a37508235963e695e640"
+      "spec_sha256": "3080fe938df62216dc4853c1163639817de3c537e0e932309642fbd3f7d8f266"
     }
   },
   "report_schema_version": 3,
@@ -2609,7 +2609,7 @@
     "country": "us",
     "schema_id": "country_spec",
     "schema_version": 1,
-    "spec_sha256": "638cd60e97a9b200f8bdfd4f92e67be986f516ac25b9a37508235963e695e640"
+    "spec_sha256": "3080fe938df62216dc4853c1163639817de3c537e0e932309642fbd3f7d8f266"
   },
   "status": "pass"
 }

packages/microcosm-build/src/microcosm/build/spec_engine/inventory_coverage.py

@@ -362,8 +362,8 @@
     # #1178: puf_support's source bytes attest CPS Social Security preservation.
     # RNG rules and the ordered QRF targets are unchanged. microcosm#104 moves
     # both again: the protocol attests microcosm.calibrate.solve.
-    "seed_map": "bdd339868ee0cec07fa3a52a91c4bd97de707c15aeb0dc224b064be9ce04eb4b",
-    "seed_protocol": "8d028354d487eadecf3bc4a950d70c95e5c44c792fe7fdf2d52c77e6d86a7886",
+    "seed_map": "41a64b8820bf161e316c49e3f2b666a247b58f6f6d50dd4395b89c8f5b1399f6",
+    "seed_protocol": "0e4c8a4ace5e10747065669a009691d0eee9179ca42b344f8912c5ca239ebde6",
     "source_manifest": "f4f8bfeb79268a4043a843b47cec1d1c0e58e9f5dd6bc23417ada95446dd7b8e",
     "take_up": "9522ce40f7dea569312dd7a7beb474e5a5afadd2f1e10cb5876534c3ef623d35",
     "tail": "ac92829c88a1a4fb6460d61190918d5d99c6c377fc8dd8f62f02b332d09bf59c",

packages/microcosm-build/src/microcosm/build/us_runtime/puf_support.py

@@ -607,6 +607,23 @@ def puf_tax_detail_tail_bound_quantiles_identity() -> dict[str, float]:
 }
 # The omitted 13.7% E17500 share represented reported Medicare Part B premiums.
 # Do not redistribute it: the engine computes its distinct singular Part B output.
+# Raw-PUF inputs of the archived partnership self-employment derivation
+# (:func:`_puf_partnership_se_net_earnings_from_raw_columns`). All eight are
+# required: a raw PUF carrying only some of them derives nothing, so the donor
+# build fails closed instead of substituting a different quantity.
+_PUF_PARTNERSHIP_SE_RAW_COLUMNS = (
+    "E30400",
+    "E30500",
+    "E00900",
+    "E02100",
+    "E25920",
+    "E25940",
+    "E25960",
+    "E25980",
+)
+# The archived pipeline's ``SE_DEDUCTION_FACTOR``: it treats E30400 and E30500
+# as already multiplied by this factor and divides it back out.
+_PUF_SE_NET_EARNINGS_FACTOR = 0.9235
 _PUF_PREDICTOR_PREFIX = "puf_predictor_"
 
 
@@ -4332,11 +4349,10 @@ def _person_source_values(
         return _numeric_array(arrays["E25980"]) - _numeric_array(arrays["E25960"])
     if output == "s_corp_income" and {"E26190", "E26180"}.issubset(arrays):
         return _numeric_array(arrays["E26190"]) - _numeric_array(arrays["E26180"])
-    if output == "partnership_self_employment_net_earnings" and {
-        "E25960",
-        "E26180",
-    }.issubset(arrays):
-        return _numeric_array(arrays["E25960"]) + _numeric_array(arrays["E26180"])
+    if output == "partnership_self_employment_net_earnings" and set(
+        _PUF_PARTNERSHIP_SE_RAW_COLUMNS
+    ).issubset(arrays):
+        return _puf_partnership_se_net_earnings_from_raw_columns(arrays)
     if output in _PUF_MEDICAL_EXPENSE_CATEGORY_BREAKDOWNS and "E17500" in arrays:
         return (
             _numeric_array(arrays["E17500"])
@@ -4363,6 +4379,56 @@ def _person_source_values(
     return None
 
 
+def _puf_partnership_se_net_earnings_from_raw_columns(
+    arrays: Mapping[str, Sequence[Any]],
+) -> np.ndarray:
+    """Derive partnership self-employment net earnings from raw PUF columns.
+
+    This ports, operation for operation, the derivation the archived
+    processed-PUF pipeline used for this leaf (the retired enhanced-CPS
+    repository at 42ed5d45c56df80d754fbe24cce21cfeb8d05cbe,
+    ``datasets/puf/puf.py`` lines 723-746), which matches ``part_se`` in Yale
+    Budget Lab Tax-Data's ``src/process_puf.R``. It is a
+    residual, not a read of a Schedule K-1 box 14 field: gross self-employment
+    earnings less the Schedule C and Schedule F amounts.
+
+    The column meanings below are the archived pipeline's own; this module does
+    not check them against the PUF documentation:
+
+    * ``E30400`` and ``E30500``: the taxpayer's and the spouse's taxable
+      self-employment income, already multiplied by 0.9235. Their sum is
+      divided by that factor to recover gross self-employment earnings.
+    * ``E00900`` and ``E02100``: Schedule C net profit or loss and Schedule F
+      farm income, subtracted from the gross amount.
+    * ``E25940 + E25980 - E25920 - E25960``: net partnership income. A record
+      where it is exactly zero has no partnership activity and gets zero.
+
+    No S corporation column (``E26180``, ``E26190``) enters the derivation. A
+    missing value counts as zero, as the archived ``fillna(0)`` did.
+    ``_numeric_array`` also turns a nonnumeric value into zero, where the
+    archived arithmetic would raise; on numeric float64 arrays the result
+    equals the archived formula bit for bit.
+
+    Args:
+        arrays: Mapping carrying every column of
+            ``_PUF_PARTNERSHIP_SE_RAW_COLUMNS`` at one common grain.
+
+    Returns:
+        Partnership self-employment net earnings, one value per input row.
+    """
+
+    raw = {
+        column: _numeric_array(arrays[column])
+        for column in _PUF_PARTNERSHIP_SE_RAW_COLUMNS
+    }
+    gross_se = (raw["E30400"] + raw["E30500"]) / _PUF_SE_NET_EARNINGS_FACTOR
+    schedule_c_f_income = raw["E00900"] + raw["E02100"]
+    has_partnership = (
+        raw["E25940"] + raw["E25980"] - raw["E25920"] - raw["E25960"]
+    ) != 0
+    return np.where(has_partnership, gross_se - schedule_c_f_income, 0.0)
+
+
 _TAX_UNIT_ROLE_COLUMN = "tax_unit_role_input"
 _TAX_UNIT_ROLES = ("HEAD", "SPOUSE", "DEPENDENT")
 

packages/microcosm-build/tests/engine_free/shared/test_spec_engine_loader.py

@@ -12,7 +12,7 @@ def test_semantic_hash_has_golden_vector_and_surface_separation(tmp_path) -> Non
     # Pin the domain separator, normalization rules, schema-set receipt, and
     # exact normative projection as one reviewable golden vector.
     assert first.spec_sha256 == (
-        "aadf7c8cf2b0e19e75d9dc1a7b1d8190e33af9ce62657c438029b504e839867d"
+        "0c0c8eea8f5d350f3a73867aebfcb82450cb606124dc85cc3772cc2eae971ddb"
     )
 
     second_root = _rich_minimal(tmp_path / "xy", note="second", store="local:b")