fix(pashub): wire the surveyed measured cylinder heat-loss factor

A PasHub survey can lodge a manufacturer's measured cylinder loss ("What is
the cylinder measured heat loss: 0.94 kWh/24 hours") — the SAP 10.2 §4
(48)-(50) declared-loss factor that must override the Table 2 V×L×VF
insulation computation. The site-notes path dropped it two ways: the
extractor looked for "Cylinder Measured Heat Loss:" (matches 0 fixtures) while
the PDF label is "What is the cylinder measured heat loss:" (92 fixtures), and
the mapper never wired `cylinder_heat_loss` even when captured. So the
calculator fell back to the age-band insulation default — a lossier cylinder
than surveyed — over-costing hot water and under-rating SAP.

Fix the extractor label and add `_pashub_cylinder_measured_heat_loss` (parses
the leading float, returns None for "Not known"), mirroring the gov-API and
Elmhurst paths that already pass `cylinder_heat_loss` through. 3 Woodmans
Court -1.33 → -0.09; LRHA cohort within-0.5 58.3% → 60.2%, MAE 0.515 → 0.496.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

backend/documents_parser/extractor.py

@@ -732,7 +732,7 @@ class PasHubRdSapSiteNotesExtractor:
             cylinder_size=self._get_in(data, "Cylinder Size:") or "",
             cylinder_volume_measured_l=cylinder_volume_measured_l,
             cylinder_measured_heat_loss=self._get_in(
-                data, "Cylinder Measured Heat Loss:"
+                data, "What is the cylinder measured heat loss:"
             ),
             insulation_type=self._get_in(data, "Insulation Type:"),
             insulation_thickness_mm=thickness_mm,

backend/documents_parser/tests/test_pashub_sap_accuracy_lrha_wave3.py

@@ -299,6 +299,28 @@ def test_mixed_glazing_bills_each_window_at_own_u() -> None:
     )
 
 
+@pytest.mark.skipif(
+    "497712549073" not in _BY_ID, reason="measured-cylinder-loss fixture not in manifest"
+)
+def test_measured_cylinder_loss_overrides_table2() -> None:
+    """Regression: a surveyed manufacturer's measured cylinder heat-loss factor
+    ("What is the cylinder measured heat loss: 0.94 kWh/24 hours" on 3 Woodmans
+    Court, 497712549073) must reach the calculator's SAP 10.2 §4 (48)-(50)
+    declared-loss branch and override the Table 2 insulation computation. Two
+    PasHub-path gaps dropped it (an extractor label mismatch — the field lodges
+    under "What is the cylinder measured heat loss:" not "Cylinder Measured Heat
+    Loss:" — and the mapper never wired `cylinder_heat_loss`), so the calculator
+    fell back to the age-band insulation default (a lossier cylinder than
+    surveyed) → hot water over-costed → SAP under-rated -1.33 → -0.09."""
+    fixture = _BY_ID["497712549073"]
+    epc = parse_site_notes_pdf(str(fixture.path), uprn=fixture.uprn)
+    assert epc.sap_heating is not None
+    assert epc.sap_heating.cylinder_heat_loss == 0.94
+    outcome = _evaluate("497712549073")
+    assert outcome.diff is not None
+    assert outcome.diff < 0.5, f"measured cylinder loss not applied: diff {outcome.diff:.2f}"
+
+
 @pytest.mark.skipif(not _FIXTURES, reason="no pashub_accuracy_lrha_wave3 fixtures/manifest")
 @pytest.mark.parametrize("deal_id", _IDS)
 def test_pashub_fixture_computes(deal_id: str) -> None:

datatypes/epc/domain/mapper.py

@@ -7111,6 +7111,17 @@ def _map_sap_heating(
             else None
         ),
         cylinder_insulation_thickness_mm=heating.water_heating.insulation_thickness_mm,
+        # A surveyed manufacturer's measured cylinder loss ("What is the
+        # cylinder measured heat loss:" → e.g. "0.94 kWh/24 hours") is the
+        # SAP 10.2 §4 (48)-(50) declared-loss factor: when present it MUST
+        # override the Table 2 V×L×VF insulation computation. Dropped, the
+        # calculator falls back to the age-band insulation default (a lossier
+        # cylinder than surveyed) → hot water over-costed → SAP under-rated.
+        # Gated on a lodged cylinder; mirrors the gov-API/Elmhurst paths that
+        # pass `cylinder_heat_loss` through.
+        cylinder_heat_loss=_pashub_cylinder_measured_heat_loss(
+            heating.water_heating.cylinder_measured_heat_loss, cylinder_present
+        ),
         water_heating_code=water_heating_code,
         water_heating_fuel=water_heating_fuel,
         immersion_heating_type=_pashub_immersion_type_code(
@@ -7207,6 +7218,25 @@ def _pashub_community_heating(
 # `_ELMHURST_CYLINDER_SIZE_LABEL_TO_SAP10`); the raw string was int-or-noned by
 # `_cylinder_volume_l_from_code`, silently skipping the Table 28 volume
 # convention (issue #1590 bug 6).
+def _pashub_cylinder_measured_heat_loss(
+    raw: Optional[str], cylinder_present: bool
+) -> Optional[float]:
+    """Parse the PasHub "What is the cylinder measured heat loss:" value (e.g.
+    "0.94 kWh/24 hours") into the SAP 10.2 §4 declared-loss factor (kWh/day).
+    Returns None when there is no cylinder, no lodged value, or a non-numeric
+    lodgement (e.g. "Not known") — in which case the calculator falls back to
+    the Table 2 insulation computation as before."""
+    if not cylinder_present or not raw:
+        return None
+    tokens = raw.strip().split()
+    if not tokens:
+        return None
+    try:
+        return float(tokens[0])
+    except ValueError:
+        return None
+
+
 _PASHUB_CYLINDER_SIZE_TO_SAP10: Dict[str, int] = {
     "Normal (90-130 litres)": 2,
     "Medium (131-170 litres)": 3,  # Table 28 Medium → fixed 160 L