fix: repair the three coordinates blocking migration 23

database/migrations.go

@@ -1646,7 +1646,38 @@ func addRegionToAddressUniqueness() error {
 				"load state boundaries first", blanks)
 		}
 
+		// Before attributing anything, repair coordinates that are not WGS84.
+		//
+		// Three rows in production could not be placed. Two held values like
+		// (1416737.65, 811329.71) -- Ohio State Plane South in US survey feet,
+		// loaded without ever being reprojected. A longitude cannot exceed 180,
+		// so these are unambiguously not degrees.
+		//
+		// The reprojection is only applied where it demonstrably fixes the row:
+		// the result must land inside the polygon of the county the row already
+		// claims to be in. That check is what makes this a repair rather than a
+		// guess -- EPSG:3735 was chosen because it is the only candidate that
+		// satisfies it, with Ohio North, and both zones in metres, all landing
+		// in Michigan, Maine or Quebec.
 		res, err := tx.Exec(`
+			UPDATE ohio_addresses a
+			SET geom = ST_Transform(ST_SetSRID(ST_MakePoint(ST_X(a.geom), ST_Y(a.geom)), 3735), 4326)
+			FROM ohio_counties c
+			WHERE (ABS(ST_X(a.geom)) > 180 OR ABS(ST_Y(a.geom)) > 90)
+			  AND c.bounds_geometry IS NOT NULL
+			  AND c.county_name ILIKE a.county
+			  AND ST_Contains(
+				c.bounds_geometry,
+				ST_Transform(ST_SetSRID(ST_MakePoint(ST_X(a.geom), ST_Y(a.geom)), 3735), 4326))
+		`)
+		if err != nil {
+			return fmt.Errorf("failed to reproject state plane coordinates: %w", err)
+		}
+		if n, _ := res.RowsAffected(); n > 0 {
+			log.Printf("Migration 23: reprojected %d row(s) from Ohio State Plane into WGS84", n)
+		}
+
+		res, err = tx.Exec(`
 			UPDATE ohio_addresses a
 			SET region = s.state_abbr
 			FROM us_states s
@@ -1694,6 +1725,31 @@ func addRegionToAddressUniqueness() error {
 				n, shorelineToleranceMeters)
 		}
 
+		// Last resort: the county the row already claims.
+		//
+		// A row can have unusable coordinates and still say where it is. The
+		// remaining production row sits at (-0.0001, 0.0001) -- Null Island,
+		// the placeholder a missing coordinate becomes -- while naming a real
+		// Ohio county. ohio_counties holds only Ohio counties, so a name match
+		// there is evidence of the state even when the point is worthless.
+		//
+		// The bad coordinate is not hidden by this: it still shows up under
+		// outside_us_bounds on /admin/data-quality, which is where a wrong
+		// location belongs rather than blocking a uniqueness migration.
+		res, err = tx.Exec(`
+			UPDATE ohio_addresses a
+			SET region = 'OH'
+			FROM ohio_counties c
+			WHERE (a.region IS NULL OR a.region = '')
+			  AND c.county_name ILIKE a.county
+		`)
+		if err != nil {
+			return fmt.Errorf("failed to attribute blank regions from county names: %w", err)
+		}
+		if n, _ := res.RowsAffected(); n > 0 {
+			log.Printf("Migration 23: attributed %d row(s) from their Ohio county name despite unusable coordinates", n)
+		}
+
 		// Whatever is left is a genuine coordinate error, not a missing state.
 		// The details go in the error rather than only the log, because that
 		// error is surfaced on /health -- so whoever has to decide what to do

services/multistate_integration_test.go

@@ -614,3 +614,80 @@ func TestShorelineAddressesAttributeToTheNearestState(t *testing.T) {
 		}
 	}
 }
+
+// Two of the three rows production could not place held coordinates like
+// (1416737.65, 811329.71) -- Ohio State Plane South in US survey feet, loaded
+// without ever being reprojected. A longitude cannot exceed 180, so those are
+// unambiguously not degrees.
+//
+// The repair is only applied where it demonstrably works: the reprojected point
+// must land inside the polygon of the county the row already claims. That check
+// is what separates a repair from a guess -- Ohio North and both zones in
+// metres put these same points in Michigan, Maine and Quebec.
+func TestStatePlaneCoordinatesAreRepairedOnlyWhenVerified(t *testing.T) {
+	db := setupMultiStateDB(t, false)
+
+	if _, err := db.Exec(`
+		CREATE TABLE ohio_counties (
+			id BIGSERIAL PRIMARY KEY,
+			county_name VARCHAR(255) NOT NULL,
+			bounds_geometry GEOMETRY(MULTIPOLYGON, 4326)
+		)`); err != nil {
+		t.Fatalf("create ohio_counties: %v", err)
+	}
+	// A box around Darke County, and one around a county the second row does
+	// not claim, so a wrong-county match cannot pass.
+	if _, err := db.Exec(`
+		INSERT INTO ohio_counties (county_name, bounds_geometry) VALUES
+		 ('Darke',  ST_Multi(ST_MakeEnvelope(-84.9,40.0,-84.3,40.4,4326))),
+		 ('Lucas',  ST_Multi(ST_MakeEnvelope(-84.0,41.5,-83.4,41.8,4326)))
+	`); err != nil {
+		t.Fatalf("seed counties: %v", err)
+	}
+
+	// Same State Plane coordinates; one claims Darke (correct), one claims
+	// Lucas (wrong -- reprojection lands in Darke, not Lucas).
+	for _, c := range []struct{ hash, county string }{
+		{"sp-right", "Darke"},
+		{"sp-wrong", "Lucas"},
+	} {
+		if _, err := db.Exec(`
+			INSERT INTO ohio_addresses (hash, house_number, street, unit, city, district, region, postcode, county, geom, full_address)
+			VALUES ($1,'1','Main Street','','Greenville','','','',$2,
+			        ST_SetSRID(ST_MakePoint(1416737.6505, 811329.7090),4326),'1 Main Street')
+		`, c.hash, c.county); err != nil {
+			t.Fatalf("seed %s: %v", c.hash, err)
+		}
+	}
+
+	if _, err := db.Exec(`
+		UPDATE ohio_addresses a
+		SET geom = ST_Transform(ST_SetSRID(ST_MakePoint(ST_X(a.geom), ST_Y(a.geom)), 3735), 4326)
+		FROM ohio_counties c
+		WHERE (ABS(ST_X(a.geom)) > 180 OR ABS(ST_Y(a.geom)) > 90)
+		  AND c.bounds_geometry IS NOT NULL
+		  AND c.county_name ILIKE a.county
+		  AND ST_Contains(c.bounds_geometry,
+		        ST_Transform(ST_SetSRID(ST_MakePoint(ST_X(a.geom), ST_Y(a.geom)), 3735), 4326))
+	`); err != nil {
+		t.Fatalf("reproject: %v", err)
+	}
+
+	var lng, lat float64
+	if err := db.QueryRow(`SELECT ST_X(geom), ST_Y(geom) FROM ohio_addresses WHERE hash = 'sp-right'`).Scan(&lng, &lat); err != nil {
+		t.Fatalf("read repaired row: %v", err)
+	}
+	if lng < -85 || lng > -84 || lat < 40 || lat > 40.5 {
+		t.Errorf("repaired row is at (%.4f, %.4f), which is not in Darke County", lng, lat)
+	}
+
+	// The row whose county does not corroborate the reprojection must be left
+	// alone rather than moved somewhere plausible-looking.
+	if err := db.QueryRow(`SELECT ST_X(geom), ST_Y(geom) FROM ohio_addresses WHERE hash = 'sp-wrong'`).Scan(&lng, &lat); err != nil {
+		t.Fatalf("read uncorroborated row: %v", err)
+	}
+	if lng < 180 {
+		t.Errorf("a row whose county does not corroborate the reprojection was rewritten to (%.4f, %.4f); "+
+			"the county check is what makes this a repair rather than a guess", lng, lat)
+	}
+}