Model/domain/sap10_calculator/docs/HANDOVER_POST_S0380_169.md

Handover — post Slices S0380.164..169

Branch: feature/per-cert-mapper-validation. HEAD 9ed003a5. Predecessor: HANDOVER_POST_S0380_164.md.

TL;DR

Six slices landed in two phases:

Phase 1 — cascade closure (.164 + .165): closed the LAST two open variants in the cascade-OK tier. solid fuel 2 ΔCO2/ΔPE −93/-1027 → ±0.0000 (Elmhurst-mirror §12.4.4 summer-immersion double-count, new §8.2 row). pcdb 1 Δ−0.011 SAP → +0.0000 (§9.4.11 boiler-interlock -5pp applied AFTER Eq D1 instead of before). The 41-variant heating- systems corpus now has its 27-variant cascade-OK tier fully EXACT on every metric (Σ|ΔSAP_c| ≈ 0.0001 ≈ floating-point noise).

Phase 2 — mapper unblocks (.166–.169): 11 of 16 blocked variants unblocked via 4 mapper-extension slices. 7 EXACT on first try (pcdb 3, electric 11/12/13/14, oil 2, oil 5); 4 land with cascade-side residuals pinned as forcing functions (oil 3/4/6, no system).

Slice	HEAD	Scope	Cascade-OK tier
S0380.164	302db131	§8.2 Elmhurst-mirror summer-immersion CO2/PE double-count. §12.4.4 back-boiler HW blend adds S_fuel × Table 12 annual electric term on top of Table 12d/e monthly cascade for dual-rate tariffs. Closures: solid fuel 2 ΔCO2/ΔPE → ±0.0000. Single-cert evidence flagged in new §8.2 row.	25/25 EXACT
S0380.165	3de52bcb	§9.4.11 boiler-interlock -5pp applied AFTER Eq D1, not before. Reciprocal Eq D1 weighting is non-linear in η; worksheet (217)m for pcdb 1 matches post-Eq-D1 form to 1e-4. New interlock_penalty_pp kwarg on _apply_water_efficiency. Closures: pcdb 1 Δ−0.011 SAP / +£0.24 / +1.33 CO2 / +5.70 PE → ±0.0000.	27/27 EXACT
S0380.166	589a8631	Bulk LPG mapper unblock. "Bulk LPG": 27 in _ELMHURST_MAIN_FUEL_TO_SAP10. Pcdb 3 EXACT first try.	28/28 EXACT
S0380.167	7901dda4	Electric storage 11-14 unblock. WEA / REA / OEA → 30. Electric 11/12/13/14 all EXACT first try.	32/32 EXACT
S0380.168	58a95472	Bio-liquid oil 2-6 unblock + Table 32 FAME 5.44 → 7.64 (deferred .131 TODO). BFD/BXE/BXF/BZC/B3C + 4 water labels. oil 2 (HVO) + oil 5 (Bioethanol) EXACT; oil 3/4 (FAME) closed 85 % of cost gap; oil 6 (B30K) carries cascade residual.	34 EXACT + 3 pinned
S0380.169	9ed003a5	"No system" unblock per §A.2.2. "NON": 30. Cascade lands with small residual pinned.	34 EXACT + 4 pinned

Extended handover suite at HEAD: 916 pass, 0 fail. Pyright net-zero (43 → 43).

Current residual state at HEAD 9ed003a5

Cascade-OK tier (38 variants, up from 25)

34 variants EXACT (|Δ| < 1e-3) on all 4 metrics. 4 variants carry pinned non-zero residuals (forcing functions):

Variant	ΔSAP_c	Δcost	ΔCO2	ΔPE	Notes
oil 3	+2.59	−£62	−14.58	−967	FAME cascade HW kWh diff
oil 4	+2.56	−£57	−13.35	−885	FAME cascade HW kWh diff
oil 6	+3.05	−£70	−240.66	−1113	B30K SH/HW kWh gap
no system	+1.18	−£27	−49.83	−562	§A.2.2 portable-electric defaults

All other 34 variants ±0.0000 on every metric.

Blocked tier (5 variants — community heating)

Unchanged. community heating 1/2/3/4/6. Deepest remaining tier: needs extractor field extraction + new community-fuel-type dispatch in the mapper + cascade integration for heat-network paths (Table 12c distribution loss, Table 12 heat-network codes 41/43/48/51/53/54).

Open fronts ranked by leverage

1. Community heating unblock — 5 variants (next-session work)

The deepest remaining work. The Elmhurst Summary §14.1 lodges:

Heating Type Space and Water Heating Community Heat Source Boilers / Combined Heat and Power / Heat pump Community Fuel Type Mains Gas / Electricity / Mineral oil / Coal / ...

Variant breakdown:

community heating 1 — COM + 301 + Boilers + Mains Gas → code 51 community heating 2 — COM + 302 + CHP + Mains Gas → code 48 community heating 3 — COM + 304 + Heat pump + Electricity → code 41 community heating 4 — COM + 302 + CHP + Mineral oil → code 48 (or 53 if not CHP) community heating 6 — COM + 302 + CHP + Coal → code 48 (or 54 if not CHP)

The work spans three layers:

1. Extractor: extract Community Heat Source + Community Fuel Type strings into a new CommunityHeating dataclass on ElmhurstSiteNotes.
2. Mapper: _resolve_community_heating_fuel_code(heat_source, fuel) dispatch helper that maps the (heat_source, fuel) pair to the correct Table 12 heat-network code.
3. Cascade: ensure the heat-network path (Table 12c DLF age-band, PCDB Table 322 records if lodged, Table 12 note (k) DHW-only half-standing) handles all 5 sub-variants correctly.

Likely 2-3 slices total — start with the extractor + mapper, then probe each variant's residual and close cascade gaps as they surface.

2. oil 3 / oil 4 (FAME) cascade HW kWh gap

Cascade HW kWh ~900 less than worksheet for oil 3/4 (FAME boilers). Probably SAP 10.2 Table 4b code 128/129 has a non-standard summer efficiency or Eq D1 path that the cascade isn't applying. Per-line worksheet walk: cascade _apply_water_efficiency output vs (219)m row.

3. oil 6 (B30K) cascade SH + HW kWh gap

Δcost −£70 on identical prices means kWh differs. Likely a different Table 4b code-126 path. Closes ~£70 / 240 CO2 / 1113 PE.

4. no system §A.2.2 cascade defaults

ΔSAP +1.18 — cascade thinks dwelling is more efficient than worksheet. Probable spec gap: §A.2.2 portable-electric defaults (responsiveness / control-type / Table 11 secondary fraction). Per- line walk on (210)m / (240) cost factor needed.

Standard slice workflow (unchanged)

1. Read spec page + identify rule (or Elmhurst worksheet pattern)
2. Probe one variant; verify diagnosis via monkey-patch / direct walk
3. Write failing AAA test (literal # Arrange / # Act / # Assert)
4. Implement helper / dispatch entry / mapper extension
5. Re-pin affected variants (DO NOT widen tolerance)
6. Run extended handover suite (command below)
7. Pyright net-zero check (git stash → pyright → git stash pop → pyright)
8. If mirroring Elmhurst against spec literal: add a row to SAP_CALCULATOR.md §8 "Elmhurst-mirrored spec divergences". The ≥2-cert rule applies unless the new divergence shares its shape with an already-documented row (S0380.164 added §8.2 under this exception with a single-cert flag).
9. Commit with spec citation + Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
10. Update project-heating-systems-corpus + MEMORY.md index

Test baseline at HEAD 9ed003a5

PYTHONPATH=/workspaces/model python -m pytest \
    backend/documents_parser/tests/test_summary_pdf_mapper_chain.py \
    backend/documents_parser/tests/test_heating_systems_corpus.py \
    backend/documents_parser/tests/test_elmhurst_extractor.py \
    backend/documents_parser/tests/test_elmhurst_end_to_end.py \
    domain/sap10_calculator/worksheet/tests/test_e2e_elmhurst_sap_score.py \
    domain/sap10_calculator/worksheet/tests/test_heat_transmission.py \
    domain/sap10_calculator/worksheet/tests/test_internal_gains.py \
    domain/sap10_calculator/worksheet/tests/test_solar_gains.py \
    domain/sap10_calculator/worksheet/tests/test_dimensions.py \
    domain/sap10_calculator/worksheet/tests/test_rating.py \
    domain/sap10_calculator/worksheet/tests/test_ventilation.py \
    domain/sap10_calculator/worksheet/tests/test_appendix_h_solar.py \
    domain/sap10_calculator/worksheet/tests/test_mev.py \
    domain/sap10_calculator/rdsap/tests/test_cert_to_inputs.py \
    domain/sap10_calculator/rdsap/tests/test_golden_fixtures.py \
    domain/sap10_calculator/tests/test_pcdb_table_322_lookup.py \
    domain/sap10_calculator/tests/test_pcdb_table_329_lookup.py \
    domain/sap10_calculator/tests/test_table_12a.py \
    --no-cov -q

Expected: 916 pass, 0 fail.

Memories to load (in order)

project-heating-systems-corpus            # HEAD 9ed003a5
feedback-sap-10-2-only-never-10-3
feedback-software-no-special-handling
feedback-spec-floor-skepticism
feedback-worksheet-not-api-reference
feedback-spec-citation-in-commits
feedback-verify-handover-claims
feedback-zero-error-strict
feedback-commit-per-slice
feedback-aaa-test-convention
feedback-e2e-validation-philosophy
feedback-abs-diff-over-pytest-approx
feedback-golden-residuals-near-zero
feedback-one-e-minus-4-across-the-board
feedback-bigger-slices-for-uniform-work
reference-unmapped-sap-code
reference-unmapped-api-code
project-oil-price-spec-divergence

What NOT to do

• Don't reference SAP 10.3 — track 10.2 deliberately.
• Don't widen pin tolerances — re-pin smaller or find the spec gap.
• Don't add empirical gates to keep cohort pins stable.
• Don't re-investigate Slices .91..169 — all settled.
• Don't add new helpers to domain/sap10_ml/ — on deprecation path.
• Don't treat ΔSAP=0.07 as "closed" — target is <1e-4 vs worksheet.

Master doc

The canonical architecture + API + validation doc lives at domain/sap10_calculator/docs/SAP_CALCULATOR.md (7 sections + §8 with .1 and .2 entries). S0380.164 added §8.2 for the §12.4.4 summer-immersion double-count.

Good luck.