Publish schema 7 geography dimensions

packages/microcosm-build/src/microcosm/build/uk/target_references.json

@@ -22,6 +22,7 @@
       "period": 2025,
       "metadata": {
         "contract_target_id": "obr.income_tax",
+        "diagnostic_variable_id": "efo_receipts",
         "measure_kind": "prepared_column"
       },
       "assertion_policy": "allow_source_projection"
@@ -40,6 +41,7 @@
       "period": 2025,
       "metadata": {
         "contract_target_id": "obr.ni",
+        "diagnostic_variable_id": "efo_receipts",
         "measure_kind": "prepared_column"
       },
       "assertion_policy": "allow_source_projection"
@@ -58,6 +60,7 @@
       "period": 2025,
       "metadata": {
         "contract_target_id": "obr.ni_employee",
+        "diagnostic_variable_id": "efo_receipts",
         "measure_kind": "prepared_column"
       },
       "assertion_policy": "allow_source_projection"
@@ -76,6 +79,7 @@
       "period": 2025,
       "metadata": {
         "contract_target_id": "obr.ni_employer",
+        "diagnostic_variable_id": "efo_receipts",
         "measure_kind": "prepared_column"
       },
       "assertion_policy": "allow_source_projection"
@@ -94,6 +98,7 @@
       "period": 2025,
       "metadata": {
         "contract_target_id": "obr.ni_self_employed",
+        "diagnostic_variable_id": "efo_receipts",
         "measure_kind": "prepared_column"
       },
       "assertion_policy": "allow_source_projection"
@@ -112,6 +117,7 @@
       "period": 2025,
       "metadata": {
         "contract_target_id": "obr.vat",
+        "diagnostic_variable_id": "efo_receipts",
         "measure_kind": "prepared_column"
       },
       "assertion_policy": "allow_source_projection"
@@ -130,6 +136,7 @@
       "period": 2025,
       "metadata": {
         "contract_target_id": "obr.fuel_duties",
+        "diagnostic_variable_id": "efo_receipts",
         "measure_kind": "prepared_column"
       },
       "assertion_policy": "allow_source_projection"
@@ -148,6 +155,7 @@
       "period": 2025,
       "metadata": {
         "contract_target_id": "obr.capital_gains_tax",
+        "diagnostic_variable_id": "efo_receipts",
         "measure_kind": "prepared_column"
       },
       "assertion_policy": "allow_source_projection"
@@ -166,6 +174,7 @@
       "period": 2025,
       "metadata": {
         "contract_target_id": "obr.sdlt",
+        "diagnostic_variable_id": "efo_receipts",
         "measure_kind": "prepared_column"
       },
       "assertion_policy": "allow_source_projection"
@@ -184,6 +193,7 @@
       "period": 2025,
       "metadata": {
         "contract_target_id": "obr.attendance_allowance",
+        "diagnostic_variable_id": "efo_expenditure",
         "measure_kind": "prepared_column"
       },
       "assertion_policy": "allow_source_projection"
@@ -202,6 +212,7 @@
       "period": 2025,
       "metadata": {
         "contract_target_id": "obr.carers_allowance",
+        "diagnostic_variable_id": "efo_expenditure",
         "measure_kind": "prepared_column"
       },
       "assertion_policy": "allow_source_projection"
@@ -220,6 +231,7 @@
       "period": 2025,
       "metadata": {
         "contract_target_id": "obr.child_benefit",
+        "diagnostic_variable_id": "efo_expenditure",
         "measure_kind": "prepared_column"
       },
       "assertion_policy": "allow_source_projection"
@@ -238,6 +250,7 @@
       "period": 2025,
       "metadata": {
         "contract_target_id": "obr.council_tax",
+        "diagnostic_variable_id": "efo_expenditure",
         "measure_kind": "prepared_column"
       },
       "assertion_policy": "allow_source_projection"
@@ -256,6 +269,7 @@
       "period": 2025,
       "metadata": {
         "contract_target_id": "obr.esa",
+        "diagnostic_variable_id": "efo_expenditure",
         "measure_kind": "prepared_column"
       },
       "assertion_policy": "allow_source_projection",
@@ -275,6 +289,7 @@
       "period": 2025,
       "metadata": {
         "contract_target_id": "obr.housing_benefit",
+        "diagnostic_variable_id": "efo_expenditure",
         "measure_kind": "prepared_column"
       },
       "assertion_policy": "allow_source_projection"
@@ -293,6 +308,7 @@
       "period": 2025,
       "metadata": {
         "contract_target_id": "obr.jobseekers_allowance",
+        "diagnostic_variable_id": "efo_expenditure",
         "measure_kind": "prepared_column"
       },
       "assertion_policy": "allow_source_projection"
@@ -311,6 +327,7 @@
       "period": 2025,
       "metadata": {
         "contract_target_id": "obr.pension_credit",
+        "diagnostic_variable_id": "efo_expenditure",
         "measure_kind": "prepared_column"
       },
       "assertion_policy": "allow_source_projection"
@@ -329,6 +346,7 @@
       "period": 2025,
       "metadata": {
         "contract_target_id": "obr.pip",
+        "diagnostic_variable_id": "efo_expenditure",
         "measure_kind": "prepared_column"
       },
       "assertion_policy": "allow_source_projection"
@@ -347,6 +365,7 @@
       "period": 2025,
       "metadata": {
         "contract_target_id": "obr.state_pension",
+        "diagnostic_variable_id": "efo_expenditure",
         "measure_kind": "prepared_column"
       },
       "assertion_policy": "allow_source_projection"
@@ -365,6 +384,7 @@
       "period": 2025,
       "metadata": {
         "contract_target_id": "obr.statutory_maternity_pay",
+        "diagnostic_variable_id": "efo_expenditure",
         "measure_kind": "prepared_column"
       },
       "assertion_policy": "allow_source_projection"
@@ -383,6 +403,7 @@
       "period": 2025,
       "metadata": {
         "contract_target_id": "obr.tv_licence_fee",
+        "diagnostic_variable_id": "efo_expenditure",
         "measure_kind": "prepared_column"
       },
       "assertion_policy": "allow_source_projection"
@@ -401,6 +422,7 @@
       "period": 2025,
       "metadata": {
         "contract_target_id": "obr.universal_credit_in_cap",
+        "diagnostic_variable_id": "efo_expenditure",
         "measure_kind": "prepared_column"
       },
       "assertion_policy": "allow_source_projection"
@@ -419,6 +441,7 @@
       "period": 2025,
       "metadata": {
         "contract_target_id": "obr.universal_credit_outside_cap",
+        "diagnostic_variable_id": "efo_expenditure",
         "measure_kind": "prepared_column"
       },
       "assertion_policy": "allow_source_projection"
@@ -437,6 +460,7 @@
       "period": 2025,
       "metadata": {
         "contract_target_id": "obr.winter_fuel_allowance",
+        "diagnostic_variable_id": "efo_expenditure",
         "measure_kind": "prepared_column"
       },
       "assertion_policy": "allow_source_projection"

packages/microcosm-build/src/microcosm/build/uk_runtime/diagnostics.py

@@ -1,12 +1,11 @@
 """Standard calibration diagnostics for UK release candidates.
 
 The shared :mod:`microcosm.calibrate.diagnostics` payload is the release
-contract: it carries the target surface, every target row, solver options, and
-the concentration scalars used by US releases.  UK needs a little more release
-evidence without changing that shared schema (and therefore without changing
-US output): the effective-sample-size fraction, shipped-weight concentration,
-zero-weight rows split by their declared support strata, and target fit by UK
-geography level.
+contract: it carries the target surface, every target row, solver options,
+schema-7 source/variable/dimension identity, and the concentration scalars used
+by US releases. UK adds the effective-sample-size fraction, shipped-weight
+concentration, zero-weight rows split by their declared support strata, and
+target fit by UK geography level.
 
 This module wraps the shared payload and places those additions under a
 separately versioned ``uk_diagnostics`` block.  The common top-level

packages/microcosm-build/tests/test_uk_diagnostics.py

@@ -577,7 +577,7 @@ def test_payload_requires_a_valid_matching_uk_registry() -> None:
             target_geography_levels=geography,
             target_registry=TargetRegistry((), country="uk"),
         )
-    with pytest.raises(ValueError, match="exactly partition"):
+    with pytest.raises(ValueError, match="does not contain compiled target row"):
         uk_calibration_diagnostics_payload(
             result,
             frame,

packages/microcosm-build/tests/test_uk_rowwise_candidate.py

@@ -399,7 +399,7 @@ def test_candidate_build_writes_calibrated_h5_and_evidence(
     assert diagnostics["metric"].unique().tolist() == ["households"]
     assert len(support) == 8
     assert past_cap["n_targets"] == 4
-    assert calibration_diagnostics["schema_version"] == 6
+    assert calibration_diagnostics["schema_version"] == 7
     uk_diagnostics = calibration_diagnostics["uk_diagnostics"]
     assert len(uk_diagnostics["weakest_families"]) == 1
     assert len(uk_diagnostics["weakest_areas_by_fit"]["bottom_by_fit"]) == 4

packages/microcosm-build/tests/test_uk_target_references.py

@@ -24,9 +24,12 @@
     TargetReferenceAuthoringConfig,
     author_target_references,
 )
+from microcosm.calibrate.diagnostics import diagnostics_payload
 from microcosm.calibrate.matrix import build_constraint_matrix
+from microcosm.calibrate.score import score_targets
 from microcosm.frame import EntitySchema, Frame, WeightKind, Weights
 from tools.generate_uk_target_references import (
+    OBR_DIAGNOSTIC_VARIABLE_BY_TARGET_ID,
     POLICYENGINE_BINDING_KEYS,
     _annual_uc_award_band_token,
     _geography_pins,
@@ -128,10 +131,16 @@ def test_uk_target_references_follow_contract_derivation_rules() -> None:
         assert reference["measure"] == expected_measure
         assert reference["family"] == target["family"]
         assert reference["period"] == 2025
-        assert reference["metadata"] == {
+        expected_metadata = {
             "contract_target_id": contract_target_id,
             "measure_kind": "prepared_column",
         }
+        diagnostic_variable_id = OBR_DIAGNOSTIC_VARIABLE_BY_TARGET_ID.get(
+            contract_target_id
+        )
+        if diagnostic_variable_id is not None:
+            expected_metadata["diagnostic_variable_id"] = diagnostic_variable_id
+        assert reference["metadata"] == expected_metadata
         # The measure is a prepared column, so the pointed-to contract binding
         # must carry what the microcosm#622 materializer needs to prepare it.
         assert (
@@ -152,6 +161,29 @@ def test_uk_target_references_do_not_bind_known_mismatched_property_amounts() ->
     ]
 
 
+def test_uk_obr_references_declare_receipts_and_expenditure_categories() -> None:
+    references = _load_uk_resource("target_references.json")["target_references"]
+    obr_references = [
+        reference
+        for reference in references
+        if reference["metadata"]["contract_target_id"].startswith("obr.")
+    ]
+
+    assert len(obr_references) == len(OBR_DIAGNOSTIC_VARIABLE_BY_TARGET_ID)
+    assert {
+        reference["metadata"]["contract_target_id"]: reference["metadata"][
+            "diagnostic_variable_id"
+        ]
+        for reference in obr_references
+    } == OBR_DIAGNOSTIC_VARIABLE_BY_TARGET_ID
+    assert not [
+        reference["name"]
+        for reference in references
+        if not reference["metadata"]["contract_target_id"].startswith("obr.")
+        and "diagnostic_variable_id" in reference["metadata"]
+    ]
+
+
 def test_uk_target_references_do_not_emit_nan_uc_payment_bands() -> None:
     resource = _load_uk_resource("target_references.json")
 
@@ -418,6 +450,7 @@ def test_uk_target_references_compile_from_real_staged_feed_rows() -> None:
     assert income_tax.period == 2025
     assert income_tax.metadata["ledger_assertion"] == "source_projection"
     assert income_tax.metadata["ledger_assertion_policy"] == ("allow_source_projection")
+    assert income_tax.metadata["diagnostic_variable_id"] == "efo_receipts"
 
     tcl_households = targets["dwp.uc.two_child_limit.households_affected"]
     assert tcl_households.value == 469_780
@@ -558,6 +591,20 @@ def test_uk_target_references_constrain_a_frame_with_prepared_columns() -> None:
     for name, target_value in zip(problem.names, problem.target_vector, strict=True):
         assert target_value == fact_values_by_name[name], name
 
+    diagnostics = diagnostics_payload(
+        score_targets(frame, registry.to_target_set()),
+        target_registry=registry,
+    )
+    obr_variables = {
+        row["name"]: row["variable"]["id"]
+        for row in diagnostics["targets"]
+        if row["source"]["id"] == "obr"
+    }
+    assert obr_variables == {
+        "obr.esa@2025": "efo_expenditure",
+        "obr.income_tax@2025": "efo_receipts",
+    }
+
 
 def _real_uk_consumer_fact_rows() -> list[dict]:
     """Public rows copied from .codex-work/consumer_facts_uk.jsonl."""