feat: harden sync flow, proxy bridge, and data normalization

tests/test_db.py

@@ -8,7 +8,7 @@ from sqlalchemy import select
 
 from iaai_scraper.core.config import Settings
 from iaai_scraper.storage.db import PersistenceService
-from iaai_scraper.storage.models import Car, Image
+from iaai_scraper.storage.models import Car, Image, SyncRun
 from iaai_scraper.storage.schemas import CarRecord, ImageRecord
 
 
@@ -116,6 +116,36 @@ class TestPersistenceServiceIntegration(unittest.TestCase):
             car = session.execute(select(Car).where(Car.origin_id == "1000")).scalar_one()
         self.assertEqual(car.origin_id, "1000")
 
+    def test_start_sync_run_marks_stale_running_runs_as_failed(self) -> None:
+        first_run_id = self.persistence.start_sync_run("lane-a")
+        second_run_id = self.persistence.start_sync_run("lane-b")
+
+        self.assertNotEqual(first_run_id, second_run_id)
+
+        with self.persistence.session_scope() as session:
+            first = session.get(SyncRun, first_run_id)
+            second = session.get(SyncRun, second_run_id)
+
+        self.assertEqual(first.status, "failed")
+        self.assertIsNotNone(first.finished_at)
+        self.assertEqual(second.status, "running")
+
+    def test_upsert_falls_back_to_origin_url_to_prevent_duplicates(self) -> None:
+        first = self._record("OLD-ID", content_hash="v1")
+        first.origin_url = "https://www.iaai.com/VehicleDetail/45089484~US"
+        self.persistence.upsert_car(first)
+
+        second = self._record("NEW-ID", content_hash="v2")
+        second.origin_url = "https://www.iaai.com/VehicleDetail/45089484~US"
+        result = self.persistence.upsert_car(second)
+
+        self.assertEqual(result["action"], "updated")
+        with self.persistence.session_scope() as session:
+            cars = session.execute(select(Car)).scalars().all()
+
+        self.assertEqual(len(cars), 1)
+        self.assertEqual(cars[0].origin_id, "NEW-ID")
+
 
 if __name__ == "__main__":
     unittest.main()