The neighbour SAP-divergence cohort was (postcode, property_type, built_form),
which mixed electric- and gas-heated dwellings. Electricity scores materially
lower in SAP than mains gas for identical fabric, so a mixed-fuel postcode
produced cross-fuel false outliers — an electric dwelling flagged only for being
electric among gas neighbours.
Add the main-heating fuel class to the cohort key (electricity variants — 29/30
+ off-peak 31-40 — collapse to one class; every other code is its own bucket).
Now every flagged divergence is a same-fuel comparison worth investigating. On
portfolio 814 this refined 20 raw divergences to 17 same-fuel ones while keeping
the genuine within-fuel outliers (e.g. the all-electric WC2B 4AW flats at SAP
26/38 vs a cohort median of 67).
Reaches the fuel via epc_main_heating_detail through the indexed property_id —
still never touches the recommendation table. Surfaced by the portfolio-814
recommendation audit (Work item B, #1388).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
PRD #1385 phase 2. 21.0.0 was the last mapper still dropping the lodged
percent_draughtproofed (top-level int) — every other path (17.0/…/21.0.1) reads
it. Dropping it understated the SAP §2 draughtproofing/infiltration credit for
21.0.0 certs.
Verification: new test 1 passed (harvested 21.0.0 sample lodges 100); corpus
6002 passed; SAP-accuracy regressions pass; pyright unchanged (39 -> 39).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
PRD #1385, building-part fabric family. The 17.0 mapper hardcoded
floor_insulation_thickness / sap_room_in_roof to None and omitted
flat_roof_insulation_thickness, all of which the 17.0 schema lodges and the
19.0/20.0/21.x paths read:
* floor_insulation_thickness: "NI" (no measured thickness) -> None so u_floor
reaches its age-band default, else pass through (mirrors the 19.0 note).
* flat_roof_insulation_thickness: pass through (was dropped).
* sap_room_in_roof: map to SapRoomInRoof(floor_area, construction_age_band),
None if absent (mirrors 19.0).
17.0-only: the 17.1 SapBuildingPart schema has none of these fields, so its None
there is correct (verified against the dataclass). roof_insulation_thickness is
intentionally left raw -- the sloping-ceiling resolve helper moves roof U-values
and is a separate follow-up.
Verification: new test 3 passed; test_mapper_corpus 6002 passed; SAP-accuracy /
RealCertExpectation regressions pass; pyright unchanged (39 -> 39, pre-existing).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
PRD #1385, shower/WWHRS family. The 17.0 and 19.0 API mappers left
sap_heating.number_baths / mixer_shower_count as None, so the water-heating
demand (cert_to_inputs) fell back to defaults instead of the lodged bath/shower
room counts. Derive them from schema.sap_heating.instantaneous_wwhrs, mirroring
the 17.1/18.0/20.0 paths:
number_baths = rooms_with_bath_and_or_shower + rooms_with_bath_and_mixer_shower
mixer_shower_count = rooms_with_mixer_shower_no_bath + rooms_with_bath_and_mixer_shower
Guarded on instantaneous_wwhrs is not None (Optional in the schema), like 17.1.
Counts, same meaning across the reduced-field family.
Verification: new test 2 passed; test_mapper_corpus 6002 passed; SAP-accuracy /
RealCertExpectation regressions pass; pyright unchanged (39 -> 39, pre-existing).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
PRD #1385, heating-controls family. The 17.0 API mapper omitted
main_heating_controls entirely, leaving the domain field None while the
19.0/20.0/21.x paths map the first lodged control via _map_energy_element. 17.0
lodges main_heating_controls (List[EnergyElement]) in all 23 local fixtures.
The mapped element is the same EnergyElement type 17.0 already carries for
main_heating / window, so this is straight parity (not new logic). Guarded on a
non-empty list -> None, matching the siblings.
Verification: new test 2 passed; test_mapper_corpus 6002 passed; recommendation
+ heating-control tests 85 passed (the domain EnergyElement field is distinct
from the subscriptable dict HeatingControlRecommender reads via the Property
wrapper, so no recommender destabilisation); SAP-accuracy regressions pass;
pyright unchanged (39 -> 39, pre-existing).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
PRD #1385, lighting family. The 17.0 and 19.0 API mappers hardcoded every
fixed-lighting bulb count to 0, dropping the lodged lighting mix that the
reference-complete 17.1/18.0/20.0 paths read. The schema lodges total +
low-energy OUTLET counts (ADR-0028); mirror the reference split — low-energy
outlets → low_energy bulbs, the remainder → incandescent — so the SAP
lighting-energy calc sees the real low-energy credit instead of assuming none.
led/cfl stay 0 (reduced-field certs don't split those); 21.0.x use a different
bulb-based lighting schema and are out of scope. The `total - low_energy`
subtraction is safe: all 37 local 17.0/19.0 fixtures have low_energy <= total
(no negative incandescent), matching the reference paths that already do this
unclamped.
Verification: new test 2 passed; test_mapper_corpus 6002 passed; SAP-accuracy /
RealCertExpectation regressions 56 passed (no pinned shifts); pyright unchanged
(39 -> 39, pre-existing).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
PRD #1385, ventilation family. The 17.0 and 19.0 API mappers dropped lodged
ventilation fields that the reference-complete 17.1/18.0/20.0 paths read:
* open_fireplaces_count -> open_chimneys_count: hardcoded 0 on both
(understating §2 infiltration).
* percent_draughtproofed: omitted on both.
* 17.0 additionally passed no sap_ventilation block, so the §2 cascade fell
back to NATURAL + its default sheltered_sides=2. Restored from the lodged
mechanical_ventilation + built_form via the existing _api_* helpers.
Semantic parity verified (not just type-equality): built_form and
mechanical_ventilation use the same gov code lists the 17.1/18.0/19.0/20.0/21.x
paths already map through these helpers; all 244 RdSAP fixtures resolve with zero
UnmappedApiCode, and the strict-coverage guard raises on any future divergence.
For the current corpus the 17.0 score-mover is sheltered_sides from built_form
(every 17.0 cert lodges mechanical_ventilation=0 -> NATURAL).
Adds a regression test asserting both paths carry the fields through
from_api_response on real certs.
Verification: new test 2 passed; test_mapper_corpus 6002 passed; SAP-accuracy /
RealCertExpectation regressions pass; pyright unchanged (39 -> 39, pre-existing).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The solid-brick LPG fixture (main_fuel_type=27, gas_connection_available=True)
was previously excluded from HHRSH by the mains_gas bug. After the fix it is
correctly offered HHRSH, so the integration test's product catalogue and
assertion are updated to match.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
One-time script (dry-run default, --apply in a transaction, idempotent) reusing the
live glazing_mix_guard so the backfill and classifier cannot drift. Maps every
genuine percentage-mix override (neither type >= 90%) off a flattened single type
onto MIXED (no overlay -> the cert's per-window glazing is kept). property_overrides
TEXT updated now; the glazing pgEnum cache write for 'Mixed glazing' is deferred
until the FE adds the member. Dry-run against the audited DB reports 647 rows.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Wires GuardedColumnClassifier(glazing_mix_guard, <llm>) into the glazing column so a
structured percentage mix deterministically resolves to MIXED (no overlay), while
uniform and varied phrasings fall through to the LLM (#1376, ADR-0042).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
A focused sibling to audit-ara-portfolio: that skill audits baselines/plans/SAP;
this one audits the *recommendations themselves* — why a measure was or wasn't
offered. Motivated by the portfolio-814 review (Khalim's HHRSH-on-community-
heating, missing-HHRSH, missing-secondary-heating-removal, and a neighbour split).
Adds:
- .claude/skills/find-weird-recommendations/SKILL.md — scan -> neighbour scan ->
live re-model deep-dive -> root-cause -> codify, with a seeded known-bug
catalogue and the query-safety rules inherited from audit-ara-portfolio.
- scripts/audit/anomalies.py: new `plan-stops-short-of-goal` HIGH check — the
default plan ends below the goal band on an unlimited-budget scenario (the
deterministic worklist for "why didn't this get recommended X"). Adds
scenario_budget to the bundle/query so budget-capped scenarios are excluded.
- scripts/audit/neighbour_divergence.py: groups a portfolio by (postcode,
property_type, built_form) and flags effective-SAP outliers vs the cohort
median. Never touches the 26m-row recommendation table, so it is safe
portfolio-wide.
- Tests for both (12 passing).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
`from_rdsap_schema_17_0` was an older copy that hardcoded a cluster of
lodged `sap_heating` fields to None while every other API path
(19.0/20.0/…) reads them from `schema.sap_heating`. PR #1380 restored the
secondary-heating pair; this restores the rest of the same block, all of
which `RdSapSchema17_0` lodges:
MainHeatingDetail: boiler_flue_type, fan_flue_present,
central_heating_pump_age, main_heating_index_number (PCDB efficiency index)
SapHeating: cylinder_thermostat, cylinder_insulation_thickness
Impact — cylinder_thermostat (12/23 local 17.0 fixtures) read as "no
thermostat", so the water-heating worksheet over-applied its ×1.3 penalty
and `recommend_cylinder_thermostat` could fire on a dwelling that already
has one; cylinder_insulation_thickness (13/23) discarded the measured value
so the cylinder-loss cascade fell back to age-band defaults;
main_heating_index_number lost the PCDB efficiency lookup.
Mirrors the 19.0 path exactly. secondary_*_type left untouched so this PR
stays independent of #1380 (no overlapping lines). Adds a regression test
asserting the whole block survives `from_api_response` on a real 17.0 cert.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Per-window assignment from a landlord aggregate is strictly worse than the cert's
existing per-window sap_windows, so a mixed override always defers — the proportion
comparison only fed a 'flag when far' branch not worth a new schema field. Realized
decision: a GlazingType.MIXED sentinel (LLM classifies mixes to it) resolves to no
overlay, preserving the cert glazing; uniform assertions still apply.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The `from_rdsap_schema_17_0` path hardcoded `sap_heating.secondary_heating_type`
and `secondary_fuel_type` to None, unlike every other API path (19.0/20.0/…),
which read them from `schema.sap_heating`. The gov EPC API *does* populate a
secondary heating SAP code for a genuine fixed secondary (e.g. 691 "electric
room heaters", 694), so the hardcode silently discarded it.
Because `recommend_secondary_heating_removal` gates purely on
`sap_heating.secondary_heating_type`, any 17.0 cert carrying a fixed secondary
would never get its removal recommendation. Surfaced while auditing portfolio
814's recommendations; 814 has 17.0 certs but none with a fixed secondary, so
the defect was latent there — this restores parity with the other paths so it
cannot bite.
Mirrors the 19.0 path exactly (secondary_fuel_type via `_api_secondary_fuel_type`).
Adds a regression guard that injects a fixed secondary into a real 17.0 cert and
asserts the code survives the mapping.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
One-time script (dry-run default, --apply in a transaction, idempotent) mapping a
flat roof's stated depth (flat: Nmm) onto the nearest Table 16 rung member so the
overlay scores it by thickness, not the age-band default. Updates property_overrides
(TEXT) immediately; the roof_type pgEnum cache writes are deferred until the FE adds
the FLAT_*MM members. Dry-run against the audited DB reports 102 rows, deferring
Flat 50/100/150 mm.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
One-time script (dry-run default, --apply in a transaction, idempotent) that maps
every party-ceiling override (another/same dwelling or premises above) currently
stored as an external roof value onto the party-ceiling member, reusing the live
roof_party_ceiling_guard so the backfill and the classifier cannot drift. Updates
property_overrides (TEXT, the modelling read path) and the roof classifier cache;
dry-run against the audited DB reports 106 rows. No FE migration — the party-
ceiling values already exist in the roof pgEnum.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Wires GuardedColumnClassifier(roof_party_ceiling_guard, <llm>) into the roof_type
column so a party-ceiling marker (another/same dwelling or premises above), with or
without a trailing depth, deterministically resolves to its party-ceiling member
(no overlay, ~0 heat loss) instead of the LLM occasionally reading it as a pitched
loft (#1376).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
