Overlay PasHub as-surveyed performance onto saved Site-Notes property 🟩

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

backend/pashub_fetcher/pashub_service.py

@@ -1,6 +1,7 @@
 import os
+from dataclasses import replace
 from datetime import datetime, timezone
-from typing import Callable, List, NamedTuple, Optional, cast
+from typing import Any, Callable, Dict, List, NamedTuple, Optional, cast
 
 from backend.app.db.connection import db_session
 from infrastructure.postgres.uploaded_file_table import (
@@ -15,11 +16,13 @@ from backend.pashub_fetcher.pashub_client import (
     DownloadedFile,
     DownloadedFiles,
     PashubClient,
+    RdSapSummary,
     UnauthorizedError,
 )
 from backend.pashub_fetcher.pashub_to_ara_trigger_request import (
     PashubToAraTriggerRequest,
 )
+from datatypes.epc.domain.epc import Epc
 from datatypes.epc.domain.epc_property_data import EpcPropertyData
 from utils.logger import setup_logger
 from utils.s3 import upload_file_to_s3
@@ -28,6 +31,31 @@ from utils.sharepoint.domna_sharepoint_client import DomnaSharepointClient
 logger = setup_logger()
 
 
+def _overlay_lodged_performance(
+    epc_data: EpcPropertyData, summary: RdSapSummary
+) -> EpcPropertyData:
+    """Fill the as-surveyed current-performance block PasHub carries but the
+    Site Notes PDF does not. Best-effort per field: values PasHub returns as
+    null are left empty. The EPC Band is derived from the SAP score via
+    `Epc.from_sap_score` rather than PasHub's opaque band code."""
+    updates: Dict[str, Any] = {}
+
+    if summary.pre_sap_rating is not None:
+        score = int(summary.pre_sap_rating)
+        updates["energy_rating_current"] = score
+        updates["current_energy_efficiency_band"] = Epc.from_sap_score(score)
+
+    if summary.pre_current_co2_emissions is not None:
+        updates["co2_emissions_current"] = summary.pre_current_co2_emissions
+
+    if summary.pre_current_energy_consumption is not None:
+        updates["energy_consumption_current"] = int(
+            summary.pre_current_energy_consumption
+        )
+
+    return replace(epc_data, **updates)
+
+
 class _FileUploadRecord(NamedTuple):
     file_path: str
     file_type: Optional[str]
@@ -101,7 +129,7 @@ class PashubService:
             upload_records = self._upload_to_s3_and_update_db(
                 downloaded.core, uprn, hubspot_deal_id, file_source
             )
-            self._save_site_notes(upload_records)
+            self._save_site_notes(upload_records, active_client, job_id)
 
             if downloaded.other:
                 self._upload_to_s3_and_update_db(
@@ -190,17 +218,44 @@ class PashubService:
 
         return upload_records
 
-    def _save_site_notes(self, upload_records: List[_FileUploadRecord]) -> None:
+    def _fetch_rdsap_summary(
+        self, active_client: PashubClient, job_id: str
+    ) -> RdSapSummary:
+        try:
+            return active_client.get_rdsap_summary_by_job_id(job_id)
+        except UnauthorizedError:
+            if active_client is not self._pashub_client:
+                raise
+            logger.info(
+                f"PasHub credentials unauthorized for RdSAP Summary on job {job_id}; "
+                "retrying with CoordinationHub credentials"
+            )
+            return self._get_coordination_client().get_rdsap_summary_by_job_id(job_id)
+
+    def _save_site_notes(
+        self,
+        upload_records: List[_FileUploadRecord],
+        active_client: PashubClient,
+        job_id: str,
+    ) -> None:
         for record in upload_records:
             if (
                 record.file_type is None
                 or FileTypeEnum(record.file_type) != FileTypeEnum.RD_SAP_SITE_NOTE
             ):
                 continue
+
+            # The RdSAP Summary carries the as-surveyed Lodged Performance the
+            # Site Notes PDF lacks. Fetch it outside the parse try/except so a
+            # missing summary fails loudly rather than persisting a Property with
+            # an incomplete Lodged Performance.
+            summary: RdSapSummary = self._fetch_rdsap_summary(active_client, job_id)
+
             try:
                 epc_data: EpcPropertyData = parse_site_notes_pdf(
                     record.file_path, uprn=record.uprn
                 )
+                epc_data = _overlay_lodged_performance(epc_data, summary)
                 with db_session() as session:
                     save_epc_property_data(
                         session, epc_data, uploaded_file_id=record.uploaded_file_id

backend/pashub_fetcher/tests/test_pashub_service.py

@@ -290,13 +290,15 @@ def test_run_uses_hubspot_deal_id_path_when_no_uprn() -> None:
 
 
 def test_run_parses_and_saves_site_notes_for_rd_sap_site_note_file() -> None:
-    mock_client = MagicMock(spec=PashubClient)
-    mock_client.get_uprn_by_job_id.return_value = None
-    mock_client.get_evidence_files_by_job_id.return_value = make_downloaded(
-        core=["/tmp/RdSAP_SiteNote_001.pdf"]
+    mock_client = _site_note_client(
+        summary=RdSapSummary(
+            pre_sap_rating=None,
+            pre_current_co2_emissions=None,
+            pre_current_energy_consumption=None,
+        )
     )
 
-    fake_epc_data = MagicMock()
+    parsed = make_epc_property_data()
     fake_session = MagicMock()
     fake_uploaded_file_id = 99
 
@@ -306,7 +308,7 @@ def test_run_parses_and_saves_site_notes_for_rd_sap_site_note_file() -> None:
         patch("backend.pashub_fetcher.pashub_service.upload_file_to_s3"),
         patch(
             "backend.pashub_fetcher.pashub_service.parse_site_notes_pdf",
-            return_value=fake_epc_data,
+            return_value=parsed,
         ) as mock_parse,
         patch(
             "backend.pashub_fetcher.pashub_service.save_epc_property_data"
@@ -320,9 +322,9 @@ def test_run_parses_and_saves_site_notes_for_rd_sap_site_note_file() -> None:
         mock_db.return_value.__enter__.return_value = fake_session
         service.run(make_request(uprn="12345"))
 
-    mock_parse.assert_called_once_with("/tmp/RdSAP_SiteNote_001.pdf")
+    mock_parse.assert_called_once_with(RD_SAP_SITE_NOTE_PATH, uprn=12345)
     mock_save.assert_called_once_with(
-        fake_session, fake_epc_data, uploaded_file_id=fake_uploaded_file_id
+        fake_session, parsed, uploaded_file_id=fake_uploaded_file_id
     )