refresh api and tests
This commit is contained in:
@@ -2,16 +2,18 @@ from __future__ import annotations
|
||||
|
||||
import unittest
|
||||
from datetime import datetime, timezone
|
||||
from unittest.mock import patch
|
||||
|
||||
from mobilede_scraper.mobile_de.models import MobileDeListing, MobileDeSearchPage
|
||||
from mobilede_scraper.mobile_de.scraper import MobileDeScraper
|
||||
|
||||
|
||||
class _FakePersistence:
|
||||
def __init__(self) -> None:
|
||||
def __init__(self, existing_ids: set[str] | None = None) -> None:
|
||||
self.upsert_calls: list[list[str]] = []
|
||||
self.upsert_records: list[object] = []
|
||||
self.finish_payload: dict[str, object] | None = None
|
||||
self.existing_ids = existing_ids or set()
|
||||
|
||||
def create_tables(self) -> None:
|
||||
return None
|
||||
@@ -20,7 +22,7 @@ class _FakePersistence:
|
||||
return 1
|
||||
|
||||
def get_existing_origin_ids(self, origin_ids: list[str]) -> set[str]:
|
||||
return set()
|
||||
return {origin_id for origin_id in origin_ids if origin_id in self.existing_ids}
|
||||
|
||||
def upsert_cars_batch(self, records):
|
||||
self.upsert_calls.append([record.origin_id for record in records])
|
||||
@@ -38,6 +40,7 @@ class _FakePersistence:
|
||||
class _FakeClient:
|
||||
def __init__(self, pages: list[MobileDeSearchPage]) -> None:
|
||||
self.pages = pages
|
||||
self.detail_calls: list[str] = []
|
||||
|
||||
def build_make_model_param(self, make_id: str, model_id: str | None) -> str:
|
||||
return make_id if not model_id else f"{make_id};{model_id}"
|
||||
@@ -46,6 +49,26 @@ class _FakeClient:
|
||||
del kwargs
|
||||
yield from self.pages
|
||||
|
||||
def fetch_detail(self, listing_id: str) -> dict[str, object]:
|
||||
self.detail_calls.append(str(listing_id))
|
||||
return {
|
||||
"shortTitle": "BMW X1",
|
||||
"subTitle": "xDrive20i",
|
||||
"make": {"localized": "BMW"},
|
||||
"model": {"localized": "X1"},
|
||||
"price": {"grs": {"amount": "19999", "currency": "EUR"}},
|
||||
"contact": {"countryCode": "DE"},
|
||||
"attributes": [
|
||||
{"tag": "firstRegistration", "value": "05/2016"},
|
||||
{"tag": "mileage", "value": "71 500 km"},
|
||||
{"tag": "category", "value": "OffRoad"},
|
||||
{"tag": "color", "value": "Серый"},
|
||||
{"tag": "wheelDrive", "value": "xDrive"},
|
||||
{"tag": "cubicCapacity", "value": "1 998 ccm"},
|
||||
],
|
||||
"images": ["https://img.example.test/1.jpg"],
|
||||
}
|
||||
|
||||
|
||||
class TestMobileDeScraperStreamingSync(unittest.TestCase):
|
||||
def test_sync_search_upserts_each_page_during_run(self) -> None:
|
||||
@@ -110,6 +133,50 @@ class TestMobileDeScraperStreamingSync(unittest.TestCase):
|
||||
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))
|
||||
|
||||
def test_sync_search_selectively_enriches_new_records_with_missing_fields(self) -> None:
|
||||
pages = [
|
||||
MobileDeSearchPage(
|
||||
url="https://example.test/page-1",
|
||||
page_number=1,
|
||||
total_results=2,
|
||||
listings=[
|
||||
MobileDeListing(
|
||||
id="1",
|
||||
url="https://example.test/1",
|
||||
title="BMW X1",
|
||||
subtitle="xDrive20i",
|
||||
raw={"attr": {"fr": "", "ml": "", "c": "", "ecol": "", "cc": ""}},
|
||||
),
|
||||
MobileDeListing(
|
||||
id="2",
|
||||
url="https://example.test/2",
|
||||
title="BMW 320",
|
||||
raw={"attr": {"fr": "09/2016", "ml": "115 000 km", "c": "EstateCar", "ecol": "Серый", "cc": "1 998 ccm"}},
|
||||
),
|
||||
],
|
||||
),
|
||||
]
|
||||
client = _FakeClient(pages)
|
||||
persistence = _FakePersistence(existing_ids={"mobile.de:2"})
|
||||
scraper = MobileDeScraper(client=client, persistence=persistence)
|
||||
|
||||
with (
|
||||
patch("mobilede_scraper.mobile_de.scraper.MOBILEDE_SELECTIVE_DETAIL_ENRICH_ENABLED", True),
|
||||
patch("mobilede_scraper.mobile_de.scraper.MOBILEDE_SELECTIVE_DETAIL_ENRICH_MAX_PER_RUN", 1),
|
||||
patch("mobilede_scraper.mobile_de.scraper.MOBILEDE_SELECTIVE_DETAIL_ENRICH_MAX_PER_PAGE", 1),
|
||||
):
|
||||
result = scraper.sync_search(max_pages=1)
|
||||
|
||||
self.assertEqual(client.detail_calls, ["1"])
|
||||
self.assertEqual(result["detail_enriched"], 1)
|
||||
self.assertEqual(result["detail_enrich_failed"], 0)
|
||||
record_by_id = {record.origin_id: record for record in persistence.upsert_records}
|
||||
self.assertEqual(record_by_id["mobile.de:1"].drive, "4WD")
|
||||
self.assertEqual(record_by_id["mobile.de:1"].engine_volume, 1998)
|
||||
self.assertEqual(record_by_id["mobile.de:1"].body_type, "SUV")
|
||||
self.assertEqual(record_by_id["mobile.de:1"].year, 2016)
|
||||
self.assertEqual(record_by_id["mobile.de:2"].engine_volume, 1998)
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
unittest.main()
|
||||
|
||||
Reference in New Issue
Block a user