S0380.199: site-notes "Roof of Room" windows → roof windows (cross-mapper parity with S0380.198)

The Elmhurst extractor crashed parsing simulated-case-6's room-in-roof
window rows: the §11 "Location" cell "Roof of Room in Roof" wraps across
the layout prefix/suffix blocks and leaked into the glazing-type phrase
("Double between 2002 Roof of Room and 2021 in Roof" → UnmappedElmhurst-
Label). Fix (`_parse_window_from_anchors`): detect the roof-of-room
location tokens, strip them from the before/after blocks so the glazing
phrase reconstructs cleanly, and set location="Roof of Room".

Mapper: `_is_elmhurst_roof_window` gains a "Roof of Room" location branch
(highest-confidence rooflight signal, above the BP-roof-type / U>3.0
gates); `_ELMHURST_ROOF_WINDOW_U_BY_GLAZING` gains "Double between 2002
and 2021" → 2.30 (case 6 lodges the already-inclined roof-window U, so
the +0.30 inclination adjustment must not double-apply).

This is the site-notes mirror of S0380.198 (API window_wall_type=4):
both paths now route room-in-roof rooflights to (27a) at the inclined U.
Validated against the case-6 P960 worksheet at abs=1e-4:
  (27)  Windows      = 22.7408 (cascade 22.7407)
  (27a) Roof Windows = 13.0375 (cascade 13.0375, EXACT)
  (31)  ext area     = 336.13

Case 6 is pinned only on the §3 window line refs (new standalone test,
not added to the section-pin `_FIXTURES`) because its DUAL main heating
(51% rads + 49% underfloor, oil) makes the §10/§12 per-system lines
non-comparable to SapResult's aggregated fields — documented in the
fixture module. Summary mirrored to Summary_001431_case6.pdf.

Suite: 2355 passed, 1 skipped. New code: 0 pyright errors.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

backend/documents_parser/elmhurst_extractor.py

@@ -801,6 +801,12 @@ class ElmhurstSiteNotesExtractor:
         "North", "South", "East", "West", "NE", "NW", "SE", "SW",
     })
     _BP_INLINE_TOKENS = frozenset({"Main"})  # "Extension" only appears as suffix
+    # A room-in-roof window (rooflight) lodges its §11 "Location" cell as
+    # "Roof of Room in Roof", which the layout preprocessor wraps onto two
+    # tokens ("Roof of Room" in the prefix block, "in Roof" in the suffix).
+    # Detected so the window routes to a roof window (worksheet (27a))
+    # and the tokens don't leak into the glazing-type phrase.
+    _ROOF_OF_ROOM_LOCATION_TOKENS = frozenset({"Roof of Room", "in Roof"})
     # The Elmhurst Summary PDF lodges each window's glazing-type as a
     # capitalised phrase like "Double between 2002" / "Double with unknown"
     # / "Single" / "Triple" / "Secondary". The first token of that phrase
@@ -1020,6 +1026,18 @@ class ElmhurstSiteNotesExtractor:
         before = [lines[j].strip() for j in range(before_start, data_idx) if lines[j].strip()]
         after = [lines[j].strip() for j in range(manuf_idx + 4, after_end) if lines[j].strip()]
 
+        # Room-in-roof windows lodge their location as "Roof of Room in
+        # Roof" (wrapped across the prefix/suffix blocks). Detect it, pull
+        # those tokens out so they don't contaminate the glazing-type
+        # phrase, and override the wall-keyed `location` with the roof-of-
+        # room marker the roof-window classifier keys on.
+        if any(
+            t in self._ROOF_OF_ROOM_LOCATION_TOKENS for t in (*before, *after)
+        ):
+            location = "Roof of Room"
+            before = [t for t in before if t not in self._ROOF_OF_ROOM_LOCATION_TOKENS]
+            after = [t for t in after if t not in self._ROOF_OF_ROOM_LOCATION_TOKENS]
+
         glazing_type, building_part, orientation = self._compose_window_descriptors(
             before=before,
             after=after,

datatypes/epc/domain/mapper.py

@@ -3837,6 +3837,11 @@ def _is_elmhurst_roof_window(
     """
     if w.glazing_type.startswith("Single"):
         return False
+    # Explicit "Roof of Room" location lodging (simulated case 6): the
+    # surveyor placed the window on the room-in-roof, so it's a rooflight
+    # regardless of BP roof type or U-value.
+    if "roof of room" in (w.location or "").lower():
+        return True
     bp_roof_type = _elmhurst_bp_roof_type(w, survey)
     if bp_roof_type is not None and bp_roof_type.startswith(
         _ELMHURST_BP_ROOF_TYPES_WITH_ROOFLIGHTS
@@ -3852,6 +3857,12 @@ def _is_elmhurst_roof_window(
 # worksheet's (27a) line. The cohort exercises only "Double pre 2002".
 _ELMHURST_ROOF_WINDOW_U_BY_GLAZING: Dict[str, float] = {
     "Double pre 2002": 3.4,
+    # Simulated case 6 rooflights: the Summary lodges the already-inclined
+    # roof-window U=2.30 for DG-2002-2021 glazing (vs 2.00 vertical for the
+    # same glazing on a wall) — the worksheet bills it on (27a) at U_eff
+    # 2.1062 (= 2.30 with the §3.2 R=0.04 curtain transform). Keyed here so
+    # the inclination adjustment isn't double-applied.
+    "Double between 2002 and 2021": 2.30,
 }
 
 

tests/domain/sap10_calculator/worksheet/_elmhurst_worksheet_001431_case6.py

@@ -0,0 +1,108 @@
+"""Mapper-driven cascade fixture for the Elmhurst P960-0001-001431
+"simulated case 6" worksheet — a DETACHED, dual-oil cousin of golden
+cert 0240 carrying ROOM-IN-ROOF WINDOWS (rooflights).
+
+Routes the Summary PDF through ElmhurstSiteNotesExtractor +
+from_elmhurst_site_notes (no hand-built EpcPropertyData) so the pin
+exercises the whole extractor + mapper + calculator pipeline.
+
+Purpose: validate S0380.198/199 ROOF-WINDOW handling against a real
+worksheet. Case 6 lodges 6 windows on the room-in-roof ("Roof of Room"
+location); the worksheet bills them on line (27a) Roof Windows at
+U_eff 2.1062 (= inclined 2.30 with the §3.2 R=0.04 curtain transform),
+NOT on (27) as vertical glazing. This is the site-notes mirror of
+0240's API `window_wall_type=4` roof windows (S0380.198).
+
+This cert surfaced two site-notes gaps fixed in S0380.199:
+- the extractor mangled the "Roof of Room in Roof" window-location cell
+  into the glazing-type phrase ("Double between 2002 Roof of Room and
+  2021 in Roof" → UnmappedElmhurstLabel); `_parse_window_from_anchors`
+  now detects + strips those tokens and marks the window roof-of-room;
+- `_is_elmhurst_roof_window` gained a "Roof of Room" location branch,
+  and `_ELMHURST_ROOF_WINDOW_U_BY_GLAZING` an entry for the
+  already-inclined "Double between 2002 and 2021" → 2.30 (so the
+  inclination adjustment isn't double-applied).
+
+SCOPE: this fixture pins only the §3 heat-transmission WINDOW line refs
+(27)/(27a)/(31) — NOT the full SapResult. Case 6 has a DUAL main heating
+system (51% radiators + 49% underfloor, oil), and `SapResult`'s
+`main_heating_fuel_kwh_per_yr` / `pumps_fans_kwh_per_yr` aggregate the
+two systems differently from the worksheet's per-system (211)/(231)
+lines, so a full SapResult pin isn't apples-to-apples. Heating is also
+SAP code 127 here vs 0240's code 130 condensing combi — so case 6 pins
+to its OWN worksheet, not 0240's register.
+
+Source: user-simulated PDFs at `sap worksheets/golden fixture
+debugging/simulated case 6/`. Summary mirrored into the tracked
+`backend/documents_parser/tests/fixtures/Summary_001431_case6.pdf`.
+
+Worksheet §3 window pin targets (P960-0001-001431, Block 1):
+- (27)  Windows       = 19.3704 (Main) + 3.3704 (Ext1) = 22.7408 W/K
+- (27a) Roof Windows  = 6.19 m² × 2.1062 = 13.0375 W/K  (the 6 rooflights)
+- (31)  Total external element area = 336.13 m²
+
+Per [[feedback-zero-error-strict]]: pins are abs=1e-4 against the PDF.
+"""
+
+from __future__ import annotations
+
+import re
+import subprocess
+from pathlib import Path
+from typing import Final
+
+from backend.documents_parser.elmhurst_extractor import ElmhurstSiteNotesExtractor
+from datatypes.epc.domain.epc_property_data import EpcPropertyData
+from datatypes.epc.domain.mapper import EpcPropertyDataMapper
+
+
+# parents[0]=worksheet/, [1]=sap10_calculator/, [2]=domain/, [3]=tests/,
+# [4]=repo root.
+_SUMMARY_PDF: Final[Path] = (
+    Path(__file__).resolve().parents[4]
+    / "backend" / "documents_parser" / "tests" / "fixtures"
+    / "Summary_001431_case6.pdf"
+)
+
+# Worksheet §3 window line refs (Block 1 — energy rating).
+LINE_27_WINDOWS_W_PER_K: Final[float] = 22.7408
+LINE_27A_ROOF_WINDOWS_W_PER_K: Final[float] = 13.0375
+LINE_31_TOTAL_EXTERNAL_AREA_M2: Final[float] = 336.13
+
+
+def _summary_pdf_to_textract_style_pages(pdf_path: Path) -> list[str]:
+    """Convert a Summary PDF into the per-page text format the
+    ElmhurstSiteNotesExtractor expects (mirror of the case-5 helper)."""
+    info = subprocess.run(
+        ["pdfinfo", str(pdf_path)], capture_output=True, text=True, check=True,
+    ).stdout
+    m = re.search(r"Pages:\s+(\d+)", info)
+    if m is None:
+        raise RuntimeError(f"Could not parse page count from {pdf_path}")
+    page_count = int(m.group(1))
+
+    pages: list[str] = []
+    for i in range(1, page_count + 1):
+        layout = subprocess.run(
+            [
+                "pdftotext", "-layout", "-f", str(i), "-l", str(i),
+                str(pdf_path), "-",
+            ],
+            capture_output=True, text=True, check=True,
+        ).stdout
+        tokens: list[str] = []
+        for line in layout.splitlines():
+            if not line.strip():
+                tokens.append("")
+                continue
+            parts = [p for p in re.split(r"\s{2,}", line.strip()) if p]
+            tokens.extend(parts)
+        pages.append("\n".join(tokens))
+    return pages
+
+
+def build_epc() -> EpcPropertyData:
+    """Route the simulated case-6 Summary through extractor + mapper."""
+    pages = _summary_pdf_to_textract_style_pages(_SUMMARY_PDF)
+    site_notes = ElmhurstSiteNotesExtractor(pages).extract()
+    return EpcPropertyDataMapper.from_elmhurst_site_notes(site_notes)

tests/domain/sap10_calculator/worksheet/test_section_cascade_pins.py

@@ -42,6 +42,7 @@ from tests.domain.sap10_calculator.worksheet import (
     _elmhurst_worksheet_000487 as _w000487,
     _elmhurst_worksheet_000490 as _w000490,
     _elmhurst_worksheet_000516 as _w000516,
+    _elmhurst_worksheet_001431_case6 as _w001431_case6,
 )
 
 
@@ -248,6 +249,35 @@ def test_section_3_line_refs_match_pdf(
     _pin(actual, expected, f"§3 {fixture_attr} {fixture_name}")
 
 
+def test_section_3_roof_windows_case6_match_pdf() -> None:
+    """§3 (27a) roof-window pin for simulated case 6 — the 6 room-in-roof
+    rooflights (window_wall_type=4 on the API side / "Roof of Room"
+    location on the site-notes side) must bill on (27a) at U_eff 2.1062,
+    not on (27) as vertical glazing. Validates the S0380.198/199 roof-
+    window routing against a real worksheet. Case 6 is pinned only on the
+    §3 window line refs (not added to `_FIXTURES`) because its dual main
+    heating system makes the §10/§12 per-system lines non-comparable —
+    see the fixture module docstring."""
+    # Arrange
+    epc = _w001431_case6.build_epc()
+
+    # Act
+    ht = heat_transmission_section_from_cert(epc)
+
+    # Assert
+    _pin(ht.windows_w_per_k, _w001431_case6.LINE_27_WINDOWS_W_PER_K, "§3 (27) case6")
+    _pin(
+        ht.roof_windows_w_per_k,
+        _w001431_case6.LINE_27A_ROOF_WINDOWS_W_PER_K,
+        "§3 (27a) case6",
+    )
+    _pin(
+        ht.total_external_element_area_m2,
+        _w001431_case6.LINE_31_TOTAL_EXTERNAL_AREA_M2,
+        "§3 (31) case6",
+    )
+
+
 # ============================================================================
 # §4 Water heating — LINE_42..LINE_65 scalar + monthly tuples
 # ============================================================================