Track sold cars

This commit is contained in:
qananasikq
2026-05-06 21:07:24 +03:00
parent 81e99a41e8
commit 519686c0b7
8 changed files with 322 additions and 84 deletions

View File

@@ -2,6 +2,7 @@
import tempfile
import unittest
from datetime import datetime, timedelta, timezone
from pathlib import Path
from sqlalchemy import select
@@ -48,6 +49,14 @@ class TestPersistenceServiceIntegration(unittest.TestCase):
],
)
@staticmethod
def _as_utc(value: datetime | None) -> datetime | None:
if value is None:
return None
if value.tzinfo is None:
return value.replace(tzinfo=timezone.utc)
return value.astimezone(timezone.utc)
def test_insert_update_and_skip_flow(self) -> None:
first = self._record("777", price=1000)
inserted = self.persistence.upsert_car(first)
@@ -122,6 +131,57 @@ class TestPersistenceServiceIntegration(unittest.TestCase):
self.assertEqual(len(cars), 1)
self.assertEqual(cars[0].origin_id, "NEW-ID")
def test_upsert_preserves_first_seen_and_reactivates_seen_car(self) -> None:
first_seen = datetime(2026, 5, 1, tzinfo=timezone.utc)
sold_at = datetime(2026, 5, 2, tzinfo=timezone.utc)
seen_at = datetime(2026, 5, 5, tzinfo=timezone.utc)
first = self._record("mobile.de:777", price=1000)
first.first_seen_at = first_seen
first.last_seen_at = first_seen
self.persistence.upsert_car(first)
with self.persistence.session_scope() as session:
car = session.execute(select(Car).where(Car.origin_id == "mobile.de:777")).scalar_one()
car.is_sold = True
car.sold_at = sold_at
updated = self._record("mobile.de:777", price=1500)
updated.first_seen_at = seen_at
updated.last_seen_at = seen_at
updated.is_sold = False
updated.sold_at = None
self.persistence.upsert_car(updated)
with self.persistence.session_scope() as session:
car = session.execute(select(Car).where(Car.origin_id == "mobile.de:777")).scalar_one()
self.assertEqual(self._as_utc(car.first_seen_at), first_seen)
self.assertEqual(self._as_utc(car.last_seen_at), seen_at)
self.assertFalse(car.is_sold)
self.assertIsNone(car.sold_at)
self.assertEqual(car.price, 1500)
def test_mark_sold_not_seen_since_sets_sold_at_to_cycle_seen_at(self) -> None:
seen_at = datetime(2026, 5, 5, tzinfo=timezone.utc)
old = self._record("mobile.de:old")
old.first_seen_at = seen_at - timedelta(days=2)
old.last_seen_at = seen_at - timedelta(days=1)
current = self._record("mobile.de:current")
current.first_seen_at = seen_at
current.last_seen_at = seen_at
self.persistence.upsert_car(old)
self.persistence.upsert_car(current)
marked = self.persistence.mark_sold_not_seen_since(seen_at)
self.assertEqual(marked, 1)
with self.persistence.session_scope() as session:
cars = session.execute(select(Car).order_by(Car.origin_id.asc())).scalars().all()
sold_map = {car.origin_id: (car.is_sold, self._as_utc(car.sold_at)) for car in cars}
self.assertEqual(sold_map["mobile.de:old"], (True, seen_at))
self.assertEqual(sold_map["mobile.de:current"], (False, None))
if __name__ == "__main__":
unittest.main()

View File

@@ -1,6 +1,7 @@
from __future__ import annotations
import unittest
from datetime import datetime, timezone
from mobilede_scraper.mobile_de.models import MobileDeListing, MobileDeSearchPage
from mobilede_scraper.mobile_de.scraper import MobileDeScraper
@@ -9,6 +10,7 @@ from mobilede_scraper.mobile_de.scraper import MobileDeScraper
class _FakePersistence:
def __init__(self) -> None:
self.upsert_calls: list[list[str]] = []
self.upsert_records: list[object] = []
self.finish_payload: dict[str, object] | None = None
def create_tables(self) -> None:
@@ -22,6 +24,7 @@ class _FakePersistence:
def upsert_cars_batch(self, records):
self.upsert_calls.append([record.origin_id for record in records])
self.upsert_records.extend(records)
return {
"inserted": len(records),
"updated": 0,
@@ -82,6 +85,31 @@ class TestMobileDeScraperStreamingSync(unittest.TestCase):
self.assertIsNotNone(persistence.finish_payload)
self.assertEqual(int(persistence.finish_payload["cars_upserted"]), 3)
def test_sync_search_applies_one_seen_at_to_all_records_in_run(self) -> None:
seen_at = datetime(2026, 5, 5, 12, 30, tzinfo=timezone.utc)
pages = [
MobileDeSearchPage(
url="https://example.test/page-1",
page_number=1,
total_results=2,
listings=[
MobileDeListing(id="1", url="https://example.test/1", title="Car 1"),
MobileDeListing(id="2", url="https://example.test/2", title="Car 2"),
],
),
]
persistence = _FakePersistence()
scraper = MobileDeScraper(client=_FakeClient(pages), persistence=persistence)
scraper.sync_search(max_pages=1, seen_at=seen_at)
records = persistence.upsert_calls
self.assertEqual(records, [["mobile.de:1", "mobile.de:2"]])
self.assertTrue(all(record.first_seen_at == seen_at for record in persistence.upsert_records))
self.assertTrue(all(record.last_seen_at == seen_at for record in persistence.upsert_records))
self.assertTrue(all(record.is_sold is False for record in persistence.upsert_records))
self.assertTrue(all(record.sold_at is None for record in persistence.upsert_records))
if __name__ == "__main__":
unittest.main()