lgtm

pipeline/test_validate_outputs.py

@@ -0,0 +1,61 @@
+from __future__ import annotations
+
+import zipfile
+
+import polars as pl
+
+from pipeline.validate_outputs import main
+
+
+def test_validates_parquet_file_and_zip(tmp_path, monkeypatch):
+    parquet_path = tmp_path / "data.parquet"
+    file_path = tmp_path / "plain.txt"
+    zip_path = tmp_path / "archive.zip"
+
+    pl.DataFrame({"value": [1]}).write_parquet(parquet_path)
+    file_path.write_text("ok\n")
+    with zipfile.ZipFile(zip_path, "w") as archive:
+        archive.writestr("data.txt", "ok\n")
+
+    monkeypatch.setattr(
+        "sys.argv",
+        [
+            "validate_outputs",
+            "--parquet",
+            str(parquet_path),
+            "--file",
+            str(file_path),
+            "--zip",
+            str(zip_path),
+            "--glob",
+            f"{tmp_path}::*.txt",
+            "--zip-glob",
+            f"{tmp_path}::*.zip",
+        ],
+    )
+
+    assert main() == 0
+
+
+def test_rejects_missing_and_empty_outputs(tmp_path, monkeypatch, capsys):
+    empty_path = tmp_path / "empty.txt"
+    empty_path.touch()
+
+    monkeypatch.setattr(
+        "sys.argv",
+        [
+            "validate_outputs",
+            "--file",
+            str(empty_path),
+            "--parquet",
+            str(tmp_path / "missing.parquet"),
+            "--glob",
+            f"{tmp_path}::*.csv",
+        ],
+    )
+
+    assert main() == 1
+    stderr = capsys.readouterr().err
+    assert "empty file" in stderr
+    assert "missing" in stderr
+    assert "no files matched" in stderr

pipeline/utils/test_fuzzy_join.py

@@ -71,3 +71,64 @@ def test_fuzzy_join_on_postcode_requires_matching_numbers():
     ).collect()
 
     assert result["right_address"].to_list() == [None]
+
+
+def test_fuzzy_join_on_postcode_rejects_low_score_same_number_matches():
+    left = pl.LazyFrame(
+        {
+            "left_address": ["1 Example Street"],
+            "left_postcode": ["AB1 2CD"],
+        }
+    )
+    right = pl.LazyFrame(
+        {
+            "right_address": ["1 Totally Different Road"],
+            "right_postcode": ["AB1 2CD"],
+        }
+    )
+
+    result = fuzzy_join_on_postcode(
+        left=left,
+        right=right,
+        left_address_col="left_address",
+        right_address_col="right_address",
+        left_postcode_col="left_postcode",
+        right_postcode_col="right_postcode",
+    ).collect()
+
+    assert result["right_address"].to_list() == [None]
+
+
+def test_fuzzy_join_on_postcode_rejects_blank_and_invalid_match_keys():
+    left = pl.LazyFrame(
+        {
+            "left_id": ["blank", "number_only", "valid"],
+            "left_address": ["   ", "10", "10 High Street"],
+            "left_postcode": ["AB1 2CD", "AB1 2CD", "AB1 2CD"],
+        }
+    )
+    right = pl.LazyFrame(
+        {
+            "right_address": ["", "10", "10 High Street"],
+            "right_postcode": ["AB1 2CD", "AB1 2CD", "AB1 2CD"],
+        }
+    )
+
+    result = (
+        fuzzy_join_on_postcode(
+            left=left,
+            right=right,
+            left_address_col="left_address",
+            right_address_col="right_address",
+            left_postcode_col="left_postcode",
+            right_postcode_col="right_postcode",
+        )
+        .sort("left_id")
+        .collect()
+    )
+
+    assert result.select("left_id", "right_address").to_dicts() == [
+        {"left_id": "blank", "right_address": None},
+        {"left_id": "number_only", "right_address": None},
+        {"left_id": "valid", "right_address": "10 High Street"},
+    ]

pipeline/utils/test_poi_counts.py

@@ -101,6 +101,33 @@ def test_custom_radius(pois):
     assert total <= 2  # at most the co-located POIs
 
 
+def test_counts_pois_across_multiple_grid_cells_within_5km():
+    """A POI around 4.8km away must not be dropped by grid candidate lookup."""
+    postcodes = pl.DataFrame(
+        {
+            "postcode": ["GRID 5KM"],
+            "lat": [51.5],
+            "lon": [0.049],
+        }
+    )
+    pois = pl.DataFrame(
+        {
+            "lat": [51.5, 51.5],
+            "lng": [0.1183, 0.1240],
+            "category": ["Park", "Park"],
+        }
+    )
+
+    result = count_pois_per_postcode(
+        postcodes,
+        pois,
+        groups={"parks": ["Park"]},
+        radius_km=5.0,
+    )
+
+    assert result["parks_5km"][0] == 1
+
+
 def test_min_distance_finds_nearest(postcodes, pois):
     """min_distance_per_postcode returns distance to closest POI per group."""
     result = min_distance_per_postcode(postcodes, pois, groups=POI_GROUPS)