The gov-API lodges secondary fuel as an enum whose value can COLLIDE with a
different same-valued RdSAP 10 Table 32 / SAP 10.2 Table 12 fuel code:
- enum 9 = "dual fuel (mineral and wood)" vs Table code 9 = LPG SC11F
- enum 5 = "anthracite" vs Table code 5 = LPG (bulk)
The main-fuel boundary already canonicalises these (`_GOV_API_COLLISION_
FUELS`), but the SECONDARY-heating cost + CO2/PE paths never did — they took
the bare same-value lookup, so a dual-fuel room heater was priced as LPG
(3.48 vs dual-fuel 3.99 p/kWh) and emitted as LPG (CO2 0.241 vs 0.087),
and an anthracite secondary as bulk LPG (12.19 vs 3.64 p/kWh). The price
under-count over-rates SAP; the CO2 over-count inflates emissions.
Fix: add enum 9 to `_GOV_API_COLLISION_FUELS` (5 and 33 were already there)
and canonicalise the secondary fuel code on both the cost
(`_secondary_fuel_cost_gbp_per_kwh`) and factor (`_secondary_fuel_code`)
paths, mirroring the main-fuel boundary. canonical_fuel_code only touches
{5,9,33}, so genuinely Table-coded secondaries (House coal 11, wood logs 20,
community fuels 30-32) are left unchanged — confirmed by a full-map audit.
Corpus: within-0.5 69.7% -> 70.2% (MAE 0.854 -> 0.845; dual-fuel-secondary
cohort 42.9% -> 49.0%, signed +0.55 -> +0.41) and CO2 MAE 0.12 -> 0.08 t/yr
(bias +0.04 -> 0.00). Ratcheted the corpus floors (within 0.70, MAE 0.85,
CO2 0.09, PE 4.0). A prior session deferred enum 9 ("direction not
understood") while the EPC PE/CO2 lens was confounded by the climate-cascade
bug (fc7c4d2d); on the corrected lens the over-rate direction is clear.
pyright not installed in this codespace (strict gate not run locally).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The SAP/EI rating is computed on UK-average weather (Appendix U Tables
U1-U3 region 0) so ratings are nationally comparable, but Appendix U
paragraph 1 (PDF p.124) requires that "other calculations (such as for
energy use and costs on EPCs) are done using local weather. Weather data
for each postcode district are taken from the PCDB". `Sap10Calculator.
calculate` ran ONE cascade (UK-average) and fed it to SAP, CO2 AND primary
energy, so every cert's EPC-displayed CO2/PE were computed on the wrong
climate. Because most of England is warmer than the UK-average, this
systematically OVER-counted heating demand on the emissions/PE outputs.
The two cascades (`cert_to_inputs` rating, `cert_to_demand_inputs`
postcode) already existed; this wires the demand cascade into the
production entry point and grafts its CO2/PE onto the rating result (SAP
unchanged). The corpus gauge's longstanding +5% CO2/PE over-estimate was
mostly this climate bug, NOT (as previously diagnosed) per-cert mapper
fidelity:
CO2 MAE 0.26 -> 0.12 t/yr (bias +0.18 -> +0.04)
PE MAE 13.6 -> 3.8 kWh/m2 (bias +9.0 -> +0.24)
SAP within-0.5 = 69.7% (rating cascade, unchanged)
Worksheet-validated to 1e-4 on simulated case 45 (heat-pump ground-floor
flat, postcode W6): the P960 prints the current dwelling twice — Block 1
on UK-average weather (SAP 60.5318, CO2 692.13) and Block 2 on postcode
weather (CO2 626.78, PE 6581.59). Both reproduce exactly. Added a tracked
case-45 Summary fixture + two-cascade cascade pin as a permanent guard,
and ratcheted the corpus CO2/PE ceilings to 0.13 / 4.2. The e2e Elmhurst
suite (Block-1 line refs) now pins the rating cascade directly; the two
Vaillant overlay snapshots refreshed to demand-cascade CO2/PE.
pyright not installed in this codespace (strict gate not run locally);
change is type-trivial (dataclasses.replace over SapResult).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
When a heat-pump cert lodges a PCDB Table 362 record, the APM override
set BOTH the space efficiency (N3.6) and the water efficiency (N3.7a)
from the heat pump unconditionally. But the PCDB η_water applies only
when the DHW is heated BY the heat pump (water-heating code "from main":
901/902/914). A separate electric immersion (WHC 903) heats the water at
100% regardless of the space system, so applying the HP's water SCOP
(187.5% × 0.6 in-use = 112.5%) under-counted the immersion's hot-water
fuel.
Gate the η_water override on the DHW-from-main codes; a separate immersion
keeps its own 100% efficiency. Space η_space still always uses the APM
value (the heat pump is the space main).
Worksheet-validated to 1e-4 on simulated case 45 (HP space + WHC-903
immersion): water fuel (62) 1893.57 -> 2130.2639, total cost (255)
619.7433, CO2 692.13 — all matching the P960 exactly; SAP 60.53 -> rounds
to the worksheet's 61. RdSAP-21.0.1 corpus unchanged (no HP+WHC903 certs
in it). Pinned in test_cert_to_inputs (immersion fuel is main-independent).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The flat floor-exposure heuristic keys on dwelling_type: a flat defaults
to has_exposed_floor=False (assuming a heated dwelling below). The
Elmhurst Summary path lodges a ground-floor flat's vertical position as a
"Ground floor" floor_type rather than the API floor_heat_loss=1 exposed
code, and the mapper can label such a flat "Top-floor flat" — so the
cascade dropped the ground floor entirely (a ground floor is in contact
with the ground and carries heat loss).
Treat a "ground floor" floor_type as a heat-loss floor, overriding the
dwelling-level suppression upward — mirroring the existing "another
dwelling below" party override downward.
Worksheet-validated to 1e-4 on simulated case 45 (a ground-floor flat
the mapper labelled "Top-floor flat"): floor (28a) 0 -> 25.38 W/K,
fabric (33) 75.63 -> 101.0104, HTC (39) 112.93 -> 145.3579, all matching
the P960 exactly; SAP 67.81 -> 62.52. RdSAP-21.0.1 corpus within-0.5
69.5% -> 69.7% (MAE 0.859 -> 0.854). Floors ratcheted. Pinned in
test_heat_transmission (ground-floor billed + party-floor suppressed).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Adds whole-dwelling property_type/built_form to EpcSimulation (folded by
apply_simulations) and maps those override components. property_type drives
party-wall heat loss + ASHP/solar/wall eligibility, so a landlord correction now
moves both the SAP calc and the measure menu; built_form has no calculator
consumer today (feeds the ML transform). Written as the landlord text value
(park-home check is text-only). Refines ADR-0032 dec-4.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Extends WallType coverage to timber/stone/system-built/cob/park-home/curtain and
adds RoofType "Pitched, N mm loft insulation" -> roof_insulation_thickness. The
"(assumed) insulated"/"partial" wall states stay deferred (ambiguous code, needs
Elmhurst validation per ADR-0032); property_type/built_form carry no SAP weight.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The deduplicated `epc.roofs[]` list cannot be indexed 1:1 against the
building parts (190/329 multi-part certs have len(roofs) != len(parts)),
so every part's `u_roof` consumed a SINGLE join of all roof descriptions.
That leaked one part's insulation state onto another: a "Flat, no
insulation" extension dragged a "Pitched, insulated (assumed)" main roof
to the uninsulated 2.30, ~3x over-stating its heat loss. 3-part certs
systematically under-rated (56% within-0.5, mean -0.79 SAP).
Partition the non-RR roof descriptions into flat vs pitched/sloping and
match each part to its own kind (`_main_roof_descriptions_by_kind`),
falling back to the global join when a part's kind has no matching entry.
Corpus cert 100010129331: roof 110.5 -> 31.3 W/K, +13.10 -> -0.05 SAP.
RdSAP-21.0.1 within-0.5 68.8% -> 69.5% (MAE 0.888 -> 0.859; PE 13.9 ->
13.6); 3-part cohort 56% -> 61%. Floors/ceilings ratcheted. Pinned in
test_heat_transmission (by_kind split + mixed-roof no-contamination).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The §5.8 Table-14 added-insulation R-value adjustment was gated to
WALL_SOLID_BRICK, so a stone (granite/sandstone) wall lodging
wall_insulation_type 1/3 ("External"/"Internal") + a thickness fell
through the §5.6 thin-wall branch and was billed at its UNINSULATED U
(e.g. sandstone 520 mm + 100 mm internal: 1.64 instead of 0.30 → ~5×
the wall heat loss). Mirror the brick insulation branch into the stone
block, feeding the RAW §5.6 U₀ into the §5.8 chain per the same rule the
brick branch and the dry-lined granite pin 000565 already follow (the
Table-6 footnote (a) 1.7 cap does not apply on the insulated path).
Corpus cert 100052159386 (sandstone 520 mm + 100 mm internal): -26.20 ->
-4.08 SAP, walls 300 -> 55 W/K. RdSAP-21.0.1 corpus within-0.5 68.6% ->
68.8% (SAP MAE 0.942 -> 0.888; PE MAE 14.3 -> 13.9; CO2 0.27 -> 0.26);
floors/ceilings ratcheted. Unit-pinned in test_rdsap_uvalues.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Close the §6.1 conservatory demand cascade per RdSAP 10 §6.1 + Table 25.
Solar gains (§6, solar_gains.py) — Table 25 note (PDF p.51): "The
orientation of windows in a conservatory is not recorded, thus solar
gains are calculated using the default solar flux (East/West orientation,
with 20° pitch for roof windows)." The glazed wall bills onto the (76)
East line (vertical, average-overshading Z); the glazed roof onto the
(82) roof-window line (20° pitch, Z=1.0), both at Table 25 g=0.76, FF=0.70.
TFA-occupancy (mapper) — §6.1: the conservatory floor area is added to the
dwelling total floor area. TFA drives occupancy → §5 internal gains + §4
hot-water demand, so the non-separated conservatory's floor area now
enters `EpcPropertyData.total_floor_area_m2` (the worksheet's (4) = 95.38
carries it). Separated conservatories (§6.2) stay excluded.
Pinned against the case-44 P960 demand cascade at abs=1e-4: (73) internal
gains 625.1759, (83) solar gains 495.8655, (95) useful gains 1079.6510,
(99) space heating per m² 89.8073 — the full §6.1 chain reproduces EXACTLY.
The whole-dwelling SAP (72.9517) / CO2 (3241.8656) are not pinned: the
case-44 Summary omits the House-Coal secondary heater (SAP 633) the P960
descriptor carries (cf. case 43), so the cascade computes no secondary —
the entire residual (+349.77 kg CO2). A Summary-input defect, independent
of §6.1; every conservatory-affected line ref is exact. Worksheet harness
stays 47/47 0-raised; corpus unchanged (API path; mirror is the next slice).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
SAP 10.2 §2 (17)-(18): a measured/design air permeability at 50 Pa from a
Blower Door test routes infiltration via `(18) = AP50/20 + (8)`, in
preference to the components-based (16) estimate. The Elmhurst extractor
read only the AP4 ("Pulse") column of §12.2, so a Blower Door result
(§12.2 "Pressure Test Result (AP50)") fell through to the structural-
infiltration default — over-counting ventilation heat loss.
Surfaced by simulated case 44 (AP50 4.50): effective air change rate was
0.81 vs the worksheet's 0.58 (+38% ventilation loss). The cascade already
supports `air_permeability_ap50` (preferred over AP4); this wires the read
end to end (extractor → ElmhurstSiteNotes → SapVentilation → cert_to_inputs).
Pinned against the case-44 P960 §2 at abs=1e-4: (18) infiltration 0.3417
(= 4.5/20 + 0.1167) and (25) Jan effective ach 0.5812. Worksheet harness
stays 47/47 0-raised.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
SAP-Schema-16.2 (datatypes/epc/domain/mapper.py):
- 16.2 is structurally an RdSAP-17.1 cert under a different name; add
_normalize_sap_schema_16_2 (field renames + defaults) and dispatch to the
tested from_rdsap_schema_17_1 mapper. uprn_100020933699 maps → SAP 71.
- Honour a "Single glazed" windows description when multiple_glazing_type="ND"
(was defaulting to double) → RdSAP-21 code 5; eng 72→71 (lodged 70).
- 4 regression tests + sap_16_2.json fixture; 0 new pyright errors.
Flat party-wall fix (domain/sap10_calculator/worksheet/heat_transmission.py):
- Full-SAP flats carry flatness in dwelling_type, not property_type, so the
party-wall default fell through to the 0.25 house value instead of the RdSAP
Table-15 flat 0.0. Add _is_flat_or_maisonette_dwelling fallback + regression
test. uprn_10093116529 80→81 (matches the cert's lodged party u_value 0).
Accuracy corpus pins (tests/domain/sap10_calculator/test_real_cert_sap_accuracy.py):
- uprn_10093116543 (SAP-17.1 gas-combi semi): engine 81 (Elmhurst 77; documented
full-SAP→RdSAP residual — measured wall/floor U + PCDB boiler vs RdSAP defaults).
- uprn_10093116529 (SAP-17.1 g/f flat): engine 81 (Elmhurst 78).
devcontainer: add poppler-utils (pdfinfo) for the documents-parser PDF fixtures.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Wire the non-separated conservatory into the §3 heat-transmission +
§1 dimensions cascade per RdSAP 10 §6.1 (PDF p.49) + Table 25 (p.51):
"The floor area and volume of a non-separated conservatory are added to
the total floor area and volume of the dwelling. Its roof area is taken
as its floor area divided by cos(20°), and wall area is taken as the
product of its exposed perimeter and its height. ... The conservatory
walls and roof are taken as fully glazed ... Glazed walls are taken as
windows, glazed roof as rooflight."
New `worksheet/conservatory.py` derives the geometry:
- height from the equivalent storey count (§6.1: 1 storey → ground-floor
room height; 1½ → ground + 0.25 + 0.5×first; etc.);
- glazed WALL → window (27) at Table 25 U (double 3.1 / single 4.8) with
the §3.2 curtain resistance (R=0.04) → U_eff 2.758;
- glazed ROOF → rooflight (27a) at Table 25 roof U (double 3.4 / single
5.3) + curtain → U_eff 2.993;
- FLOOR → (28a) via BS EN ISO 13370 as an uninsulated SOLID ground floor
with 300 mm walls (§5.12, spec p.43), exposed perimeter = glazed
perimeter → U 0.89;
- glazed wall + roof + floor areas join (31)/(36); the fully-glazed
structure walls/roof add nothing (the glazing IS the window/rooflight).
`dimensions_from_cert` adds the conservatory floor area to TFA (4) and
floor area × height to volume (5) (feeds ventilation (8)), without making
it a storey (avg storey height for §2 infiltration is unchanged).
Pinned against the simulated case-44 P960 §3 at abs=1e-4 — every line ref
EXACT: (4) 95.3800, (5) 257.1630, (27) 96.1169, (27a) 38.2201, (28a)
21.4164, (29a) 35.5852, (30) 7.4688, (31) 294.2900, (33) 207.3274,
(36) 23.5432. The remaining whole-dwelling SAP/CO2 gap is the §6 solar
gains, closed in the next slice. Worksheet harness stays 47/47 0-raised.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The measures a run considers should come from the Scenario, not a CLI flag.
The live scenario table persists exclusions only (no inclusions column), as a
Postgres text-array of exact MeasureType values.
- Scenario gains `exclusions: frozenset[MeasureType]` + `considered_measures()`
(all measures minus the excluded ones, or None when none are excluded).
- ScenarioModel.to_domain parses the `{a,b,c}` exclusions array into
MeasureTypes, raising on a token that is not an exact MeasureType value
(no high-level category expansion), per the strict-enum convention.
- ModellingOrchestrator._plan_for derives the allowlist from the Scenario's
exclusions, combined (intersection) with any explicit considered_measures
override via the new `combine_considered_measures`.
- run_modelling_e2e sources the allowlist from the Scenario; --measures /
--exclude-measures become optional overlays (e.g. the technical
secondary_heating_removal exclusion the catalogue cannot yet stock).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The ASHP bundle is priced from the rate sheet (ADR-0025); the catalogue
row is read only for its material id, which is nullable end-to-end. The
live `material` catalogue has no `air_source_heat_pump` row, so
`products.get` raised `ValueError: no active product` and aborted every
ASHP-eligible property.
Add `ProductNotFound(ValueError)` + a concrete `ProductRepository
.get_optional`, raise the typed error from both repos, and have
`_ashp_option` look the row up optionally — a missing row now yields an
ASHP Option with `material_id=None` rather than crashing.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Two review points from @dancafc:
1) Rename the `Comparable` dataclass → `ComparableProperty` (it models one
comparable *property*; the collection stays `ComparableProperties`). Applied
across domain, repositories, orchestration, harness, scripts, and tests with a
word-boundary rename so `ComparableProperties` is untouched.
2) Move `PredictionTarget` out of comparable_properties.py into prediction_target.py
(where `PredictionTargetAttributes` + `build_prediction_target` already live).
comparable_properties.py now imports it; no import cycle (prediction_target no
longer depends on comparable_properties). Importers updated.
92 tests pass across the touched suites; pyright strict clean.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The gov API lodges a manufacturer's declared cylinder loss factor
(kWh/day) in `sap_heating.cylinder_heat_loss`, in which case it leaves
the cylinder volume / insulation type / thickness None. That field was
undeclared on the 21.0.x schemas, so `from_dict` dropped it — then
`_cylinder_storage_loss_override` hit its insulation-None / volume-None
guards and returned None, dropping the §4 storage loss ENTIRELY. The
dwelling over-rated (the declared loss is typically ~1.5 kWh/day ≈
550 kWh/yr).
SAP 10.2 §4 branch a) (PDF p.136): when the declared loss factor is
known, storage loss (50) = (48) declared loss × (49) Table-2b
temperature factor — replacing the Table 2 V×L×VF computation.
- declare `cylinder_heat_loss` on RdSapSchema21_0_0/21_0_1.SapHeating +
EpcPropertyData.SapHeating; thread through the 21.0.x mappers.
- `cylinder_storage_loss_monthly_kwh` gains `declared_loss_kwh_per_day`:
when set, combined_55 = declared × TF (volume/insulation unused).
- `_cylinder_storage_loss_override` resolves the declared loss BEFORE the
insulation/volume guards (the gov omits those when the loss is lodged).
12 /tmp certs carry it (mean |err| 3.00 -> 2.51; the clean ones close
hard, e.g. 2360 2.65 -> 0.30, 0245 2.25 -> 0.53). Corpus within-0.5
67.0% -> 67.3% (MAE 1.025 -> 1.020); /tmp 71.2% -> 71.4% (0.889 ->
0.882). Worksheet harness 47/47; regression = only the 3 pre-existing
fails; pyright net-zero.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Roofs lodged insulated at rafters carry their thickness in a DEDICATED
gov-EPC API field, `rafter_insulation_thickness` (e.g. "225mm"), while
`roof_insulation_thickness` stays None (rafters aren't loft joists). That
field was undeclared on the 21.0.x schemas, so `from_dict` silently
dropped it — the rafter certs only *looked* redacted (roof EER 2-4 =
insulated, yet no thickness), and the cascade fell to the Table 18 col (2)
unknown default (2.30), badly under-rating them.
- declare `rafter_insulation_thickness` on RdSapSchema21_0_0/21_0_1 +
EpcPropertyData.SapBuildingPart (mirrors the existing
sloping_ceiling_/flat_roof_insulation_thickness dropped-field handling).
- thread it through `from_rdsap_schema_21_0_0/21_0_1` (older schemas get
None via getattr).
- `heat_transmission` prefers `rafter_insulation_thickness` over
`roof_insulation_thickness` when the part is at-rafters, so the measured
RdSAP 10 §5.11.2 Table 16 column (2) row applies (225 mm → 0.25).
Completes the rafters roof fix: with the real thickness read, the rafter
certs are recovered rather than over-stated — cert 3100-8675-0922-8628
(band E, rafters 225mm) +8.93 → +0.43 SAP. Corpus within-0.5 67.0%
(MAE 1.025) and /tmp 71.2% (MAE 0.889) — both NET ABOVE the pre-rafters
baseline (66.9% / 70.6%). Worksheet harness 47/47; regression = only the
3 pre-existing fails; pyright net-zero.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
`u_roof` only implemented the joist column, so roofs lodged insulated at
rafters (`roof_insulation_location == 1`) were mis-billed at the joist U
on both the API and Summary paths — under-stating loss, over-rating SAP.
RdSAP 10 §5.11.2 Table 16 (spec p.42-43) gives a distinct "insulation at
rafters" column (2): the rafter cavity is shallower than a loft void, so
the same depth yields a higher U (200 mm: rafters 0.29 vs joists 0.21).
§5.11 Table 18 (p.45) likewise carries a rafters column (2) for unknown /
as-built thickness (footnote (1): "The value from the table applies for
unknown and as built") — band A-D = 2.30, E = 1.50, F = 0.68, diverging
from the joist column's 100 mm-equivalent 0.40 default (footnote (4)).
- add `_ROOF_RAFTERS_BY_THICKNESS` (Table 16 col 2) + `_ROOF_RAFTERS_BY_AGE`
(Table 18 col 2) to rdsap_uvalues; `u_roof` selects them via a new
`insulation_at_rafters` flag (ignored for flat / sloping-ceiling roofs).
- `heat_transmission` derives the flag PER BUILDING PART from
`roof_insulation_location` (gov-API int 1 / Summary "R Rafters"), which
also fixes the multi-part dedup-roof-join problem: each part's own
location now drives its U, replacing the unattributable joined
`epc.roofs[]` description.
Worksheet-validated to 1e-4: simulated case 41 (4-bp — Ext1 rafters 200mm
→ 0.29, Ext3 rafters As-Built band F → 0.68; roof total 24.8350) and case
42 (6 variants — rafters 50mm → 0.88, rafters unknown band C → 2.30,
joists/none unchanged). Case 40 stays exact (roof 35.340, total 441.1606);
worksheet harness 47/47.
Corpus within-0.5 66.9% → 66.5% (gates 0.65/1.08 hold) — a spec-correct
shift, NOT a regression: all 15 corpus rafter certs carry redacted (None)
thickness yet lodge roof EER 2-4 (insulated), so the open API blanked a
specified thickness and the spec's unknown-rafter 2.30 default correctly
over-states them. Recovery needs a roof-EER→thickness inference on the
API path (follow-up), not a change to the U-table.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
README at domain/epc_prediction/README.md — the flow diagram, where each piece
lives, links to the ADRs/CONTEXT/handover/migration note, and a runnable test
command. The team's entry point.
tests/e2e/test_epc_prediction_e2e.py — the whole gap-fill flow against the REAL
Postgres Unit of Work + EPC/Property repositories + EpcComparablePropertiesRepository
+ EpcPrediction, with only the three external HTTP clients faked (EPC API,
geospatial S3, Solar). Proves: EPC-less Property → Ingestion predicts from its
postcode cohort → persists to the predicted slot → reloaded Property resolves
effective_epc via source_path == "predicted". The canonical "see it in action".
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
build_prediction_target assembles an EPC-less Property's PredictionTarget from
its identity (postcode), resolved coordinates, and Landlord-Override attributes
(property_type / built_form / wall_construction). The eligibility GATE: a Property
whose property_type is unknown returns None — never sized from a mixed-type
cohort (ADR-0031). property_type is the hard cohort filter.
The override attributes are read through a PredictionTargetAttributesReader port
(stub seam) — the real adapter (a read over property_overrides) is being built
separately by the team; ingestion wiring depends on the abstraction and tests
substitute a fake. 2 tests (assembly + gate); pyright strict clean.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Add a `predicted_epc` slot to the Property aggregate and a "predicted" branch to
SourcePath / source_path / effective_epc (ADR-0031 decisions 1+3). A
neighbour-synthesised EpcPropertyData resolves as the Effective EPC ONLY when
there is neither a lodged EPC nor Site Notes — a real source always wins
(prediction is last-resort gap-fill). The slot is distinct from `epc` so a
predicted picture coexists with any lodged one (provenance is structural, not a
flag on EpcPropertyData); downstream consumers are untouched.
3 tests: predicted resolves when sole source; lodged EPC wins over predicted;
Site Notes win over predicted. 10/10 green, pyright strict clean.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Size the predicted dwelling from the geo-proximity-weighted median of the
cohort's floor areas rather than the plain median: homes built together share a
footprint, so a nearer neighbour's area should count for more (the same street
signal #1227 already wired into age / wall / glazing). Reuses `_geo_weights` and
adds `_weighted_median`, which reduces exactly to `statistics.median` under
uniform weights (geo off / no target coordinates) — including the even-count
midpoint average — so the MAD-minimising guarantee is preserved.
Measured over the 514-target SAP-10.2 corpus (leave-one-out):
floor_area MAE 10.48 -> 9.73 m² MAPE 13.2% -> 12.2%
Re-baselines the n=36 fixture floor_area ceiling 11.8983 -> 12.0378 (a method
change, not a loosening; the small fixture subset moved +0.14 the other way as
sample noise while the population improved decisively). The ceiling still pins
the new deterministic value exactly, so the tighten-only ratchet resumes.
Investigation ruling out the adjacent floor-area levers (kept in the follow-up):
lowering minimum_cohort (9.78-10.03, worse), hard same-form filter (10.19),
mean instead of median (10.68), constant bias correction (10.47),
extension-conditioning (oracle 9.50, not worth the misclassification cost) and
room-in-roof conditioning/additive (RiR is a confound for large multi-part
outliers — RiR area is only ~21% of total, and the increment breaks the homes
already predicted exactly). Remaining cohort lever is built-form soft-weighting,
gated on a denser corpus.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Reduced-field window U: heat_transmission derived the synthesised-window
raw U from u_window(all None) -> the 2.5 placeholder regardless of glazing.
Now routes the (uniform) glazing_type code through u_window (RdSAP Table 24)
so e.g. double pre-2002 reads 2.8, not 2.5. Only the pre-SAP10 reduced-field
path is affected (21.0.1 certs carry per-window U upstream) — the RdSAP-21.0.1
corpus gauge is unchanged at 66.9% within-0.5.
test_real_cert_sap_accuracy: pin uprn_10002468137 (RdSAP-17.1, all-electric
storage heaters) at SAP 61, validated against Elmhurst on identical inputs
(dual off-peak immersion, 110 L cylinder, 2 baths). Our engine reproduces
Elmhurst's fuel cost to the penny; lodged 55 is the old SAP-2012 schema.
Tooling to grow the accuracy corpus:
- scripts/fetch_real_life_epc_sample.py — capture a cert by UPRN into the corpus.
- scripts/compare_epc_paths.py — diff gov-API vs Elmhurst-summary EpcPropertyData
and run both through the engine, localising mapper vs calculator differences.
- skill validate-cert-sap-accuracy — the end-to-end loop (capture -> Elmhurst
inputs -> human builds -> compare -> reconcile -> pin in the test).
- skill epc-to-elmhurst-rdsap-inputs reference: corrected immersion (code 1=dual),
cylinder size (code 2 = Normal/110 L), and bath-count (WWHRS sub-tab) mappings.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Folds a haversine distance kernel into the categorical-mode weighting so a
nearer neighbour counts for more — applied ONLY to the components that showed
a clear distance signal in the corpus pre-check (age band, wall + floor
construction, glazing: homes built/retrofitted together cluster). Roof
construction showed no decay and is excluded; heating keeps its coherent
donor. Predictor stays pure: weights come from target.coordinates vs each
Comparable.coordinates (resolved at the boundary); geo is OFF when the target
has no coords, neutral for a neighbour with none.
Scale chosen on the harness: _GEO_SCALE_KM=0.1 is the gate-safe optimum
(0.05 lifts the corpus more but regresses fixture floor_construction).
Corpus (150pc/514, geo off->on): age 0.564->0.572, age_pm1 0.841->0.847,
wall 0.902->0.912, floor_con 0.786->0.796, glazing 0.667->0.673; roof
unchanged. Fixture: glazing 0.5278->0.5833 (floor ratcheted), all else held.
Refactored recency into a reusable _recency_weights vector composed via
_combine, so similarity/recency/geo factors multiply uniformly. Fixture ships
a committed _coordinates.json (OGL OS OpenData; build script carries it from
the corpus sidecar on rebuild) so the gate exercises geo without S3.
This is the per-component method applied to geography ([[feedback_per_component_best_method]]).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Adds coordinates: Optional[Coordinates] to Comparable and PredictionTarget
(data carriers — the pure predictor stays IO-free), and wires load_corpus to
read an optional _coordinates.json sidecar ({uprn: [lon, lat]}) and populate
each Comparable from its cert's uprn; iter_predictions threads the held-out
target's coordinates through. Absent sidecar -> geo-weighting stays off (no
behaviour change yet — weighting lands next slice). fetch_corpus_coordinates
now writes the sidecar into the corpus dir. load_corpus populates 99% of
corpus comparables.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Adds roof_insulation_thickness_pm1 (mirrors construction_age_band_pm1, issue
#1222): adjacent RdSAP thickness buckets (0/NI,12mm..400mm+) carry near-
identical roof U-values, so an off-by-one bucket is a SAP-neutral hit. 'ND'
(no-data) is off the ordered scale, so only an exact match counts there.
Honest measurement of SAP-relevant roof-insulation quality.
Corpus (150pc/514): exact 49.3% -> +/-1 53.7% (the misses are often multiple
buckets or ND, so the band gain is smaller than age's). Fixture: exact ==
+/-1 (0.4118) — its misses are all >1 bucket; gate floor added at 0.4118.
Also fixes two pre-existing pyright errors in the touched test file
(_epc main_fuel_type/main_heating_control were Optional but the
MainHeatingDetail attributes are non-optional Union[int, str]).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Heating sub-fields can't be field-moded without breaking system coherence,
so the whole SapHeating cluster is now copied as a unit from a single
coherent donor rather than inherited from the structural template: the
neighbour matching the cohort's modal heating signature (main fuel +
category + cylinder presence), most recent among the matches (recent cert =
current system). Including cylinder presence in the signature is load-bearing
— it protects has_hot_water_cylinder + cylinder_insulation (a bare fuel+cat
signature regressed them).
Corpus (150pc/514): heating_main_control 66.3 -> 73.9% (+7.6, the target),
main_fuel 92.8 -> 96.9, category 90.7 -> 95.7, water_fuel 92.8 -> 96.3,
water_code 88.5 -> 95.3, has_cylinder 81.1 -> 89.7, secondary 36.2 -> 42.0.
SAP MAE vs lodged 7.08 -> 6.00 (calculator floor 1.57). cylinder_insulation
-13.6 corpus (tiny-n) but +33pp on the fixture; AC requires control up +
fuel/category hold + SAP not worsened, all met.
Gate (36-target fixture): zero regression; ratcheted main_category
0.8889->0.9444, main_control 0.7500->0.8056, water_fuel 0.9167->0.9722,
water_code 0.8889->0.9444, cylinder_insulation_type 0.1667->0.5000. This is
the per-component heating method ([[feedback_per_component_best_method]]):
coherent donor, never field-mode.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Per-component method: glazing type is now the recency-weighted cohort mode
applied to every predicted window, rather than copied from the template.
Glazing is retrofitted over a dwelling's life (single -> double), so a
recent neighbour reflects the current state — same family as roof-insulation
thickness. Recency is the CORRECT weighting here: plain moding regressed the
fixture (-5.6pp) and was previously reverted; similarity weighting also
regressed it; recency improves BOTH (window geometry stays on the template,
only the glazing categorical moves).
modal_glazing_type: corpus (150pc/514) 60.7 -> 66.7% (+6.0pp); fixture
0.5000 -> 0.5278 (floor ratcheted up). Heating, geometry residuals and all
other components unchanged. Refactored _recency_weighted_mode to a reusable
_recency_weighted_choice(value_of) shared by roof insulation + glazing.
Closes the #1223 per-component approach: floor-area (median estimate) +
glazing (recency) shipped as distinct best-fit methods rather than a global
recency template, which would have disturbed the coherence-coupled heating
cluster.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Per-component method, not a global template change: the predicted floor
area is now the cohort median (the MAD-minimising point estimate of the
target's size) rather than whichever structural template's own area. The
calculator derives heat loss from building-part geometry, not this scalar,
so decoupling them is safe and the scalar becomes a better size estimate.
floor_area mean|.|: corpus (150pc/514 targets) 10.62 -> 10.48; fixture
12.2175 -> 11.8983 (ceiling ratcheted down). No other component moves.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The HHR-storage HeatingOverlay (ADR-0024) added an off-peak electric
immersion cylinder but never set `immersion_heating_type`, so the overlaid
cert left it None. The calculator then could not resolve `immersion_single`
for the SAP 10.2 Table 13 HW high-rate split and billed hot water 100% at
the off-peak low rate — £127.41 vs the relodged after-cert's £169.39,
overstating the overlay's SAP by +1.26 (CO2/PE matched, isolating it to the
HW cost path).
Add `immersion_heating_type` to HeatingOverlay, route it through
`_fold_heating` (it lives on `sap_heating`), and set it to 1 (single
off-peak immersion) on the HHR overlay to match the relodged reference.
Closes both `test_hhr_storage_overlay_reproduces_the_relodged_after_*`
cascade pins (electric-storage and no-system befores share the after).
Pre-existing failure (present before this branch's recent commits), outside
the handover regression gate. Full modelling suite 220 pass, pyright net-
zero.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
ADR-0029 decision 5: survivors were treated equally; now each neighbour's
vote in the cohort mode decays with its distance from the cohort's physical
centre (floor area from the median, age band from the modal band), so the
mode leans on the most representative neighbours instead of being swayed by
size/era outliers. Scales (size 20 m^2, age weight 0.5) chosen on the
validation corpus; the tight size kernel is load-bearing (looser scales
regress floor_insulation on the fixture).
Corpus (181 SAP-10.2 targets): wall_insulation 83.4->86.2%,
roof_construction 86.2->87.3%, floor_construction 78.8->81.2%,
floor_insulation 92.9->94.1%; net +7.5pp gained vs -1.1pp (two 1-cert dips,
both held on the fixture). Geometry/residuals untouched (template unchanged).
Gate (36-target fixture): zero regression across all 24 floors/ceilings;
ratcheted wall_insulation_type 0.7778->0.8333, floor_construction
0.7500->0.8125, floor_insulation 0.9062->0.9375. Dead _mode/_int_mode
removed (superseded by the weighted variants).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Adds PredictionConfidence (cohort size + per-component agreement = the
modal value's share among neighbours that lodge one) and
EpcPrediction.confidence(), a compute-only signal so downstream can flag
low-confidence components (ADR-0029 open item: 'confidence signal').
Sanity check on the 40-postcode corpus (1068 component predictions):
agreement is strongly predictive of correctness — pooled hit-rate 21.9%
(<0.5) / 46.7% (0.5-0.7) / 73.6% (0.7-0.9) / 95.5% (>=0.9); point-biserial
corr(agreement, correct) = 0.582. Cohort size tracks too (<6 -> 68.4%,
>=20 -> 96.0%). Surfacing / persistence is a separate HITL follow-up.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
