added tests

backend/ordnanceSurvey/main.py

@@ -153,7 +153,7 @@ def handler(body: dict[str, Any], context: Any, local: bool = False) -> None:
 
     # Process each postcode group at a time
     for postcode, group in grouped:
-        print(f"Processing postcode: {postcode} ({len(group)} rows)")
+        logger.info(f"Processing postcode: {postcode} ({len(group)} rows)")
         valid_group = AddressMatch.is_valid_postcode(postcode)
         if not valid_group:
             logger.warning(f"Postcode {postcode} is invalid, skipping")

backend/ordnanceSurvey/tests/test_address_lookup.py

@@ -0,0 +1,96 @@
+"""
+Sanity checks for the ordnanceSurvey postcode-lookup + address-match flow.
+
+Mirrors the per-row logic in `backend.ordnanceSurvey.main.handler`:
+  1. Call OS Places by postcode
+  2. Flatten to a DataFrame of candidate addresses
+  3. Score each candidate against the user input with AddressMatch.score
+  4. Pick the best match (address, UPRN, score)
+
+Hits the live OS Places API, so the test is skipped when no real API key
+is configured.
+"""
+
+from __future__ import annotations
+
+import logging
+from dataclasses import dataclass
+
+import pandas as pd
+import pytest
+
+from backend.app.config import get_settings
+from backend.ordnanceSurvey.helpers import (
+    lookup_os_places,
+    os_places_results_to_dataframe,
+)
+from backend.utils.addressMatch import AddressMatch
+
+logger = logging.getLogger(__name__)
+
+
+@dataclass(frozen=True)
+class AddressCase:
+    user_input: str
+    postcode: str
+    expected_uprn: str
+
+
+CASES: list[AddressCase] = [
+    AddressCase("3 College Grove Road", "WF1 3RL", "63025389"),
+    AddressCase("14 Musuem Street", "WA1 1HU", "100012894889"),
+    AddressCase("7 Hedgerow Walk", "WV8 1UW", "100071164800"),
+]
+
+
+@pytest.fixture(scope="module")
+def os_api_key() -> str:
+    key = get_settings().ORDNANCE_SURVEY_API_KEY
+    if not key or key in {"changeme", "test"}:
+        pytest.skip("ORDNANCE_SURVEY_API_KEY not configured for live OS Places call")
+    return key
+
+
+@pytest.mark.integration
+@pytest.mark.parametrize(
+    "case",
+    CASES,
+    ids=[f"{c.user_input} [{c.postcode}]" for c in CASES],
+)
+def test_address_lookup(case: AddressCase, os_api_key: str) -> None:
+    response = lookup_os_places(case.postcode, os_api_key)
+    assert response.get("status") == 200, f"OS Places failed: {response}"
+
+    candidates = os_places_results_to_dataframe(response["data"])
+    assert not candidates.empty, f"No candidates returned for {case.postcode}"
+
+    scores = candidates["ADDRESS"].apply(
+        lambda addr: AddressMatch.score(case.user_input, addr)
+    )
+    best_idx = scores.idxmax()
+    best_score = float(scores[best_idx])
+    best_address = candidates.at[best_idx, "ADDRESS"]
+    best_uprn = candidates.at[best_idx, "UPRN"]
+
+    logger.info(
+        "input=%r postcode=%s -> uprn=%s score=%.4f address=%r",
+        case.user_input,
+        case.postcode,
+        best_uprn,
+        best_score,
+        best_address,
+    )
+    # Also print so the result is visible without --log-cli.
+    print(
+        f"\n[{case.postcode}] {case.user_input!r}\n"
+        f"  best_address = {best_address!r}\n"
+        f"  best_uprn    = {best_uprn}\n"
+        f"  best_score   = {best_score:.4f}"
+    )
+
+    assert best_score > 0, f"No viable match for {case.user_input} ({case.postcode})"
+    assert pd.notna(best_uprn), "Expected a UPRN on the best match"
+    assert str(best_uprn) == case.expected_uprn, (
+        f"UPRN mismatch for {case.user_input} ({case.postcode}): "
+        f"got {best_uprn}, expected {case.expected_uprn}"
+    )

pytest.ini

@@ -15,6 +15,7 @@ testpaths =
     backend/export/tests
     backend/magic_plan/tests
     backend/onboarders/tests
+    backend/ordnanceSurvey/tests
     backend/pashub_fetcher/tests
     datatypes/epc/domain/tests
     datatypes/epc/schema/tests