Add checkpoint logic and tests

This commit is contained in:
qananasikq
2026-04-15 11:10:32 +03:00
parent d9111be2d2
commit acbb045b6e
8 changed files with 432 additions and 175 deletions

View File

@@ -11,6 +11,7 @@ class _FakePage:
def __init__(self, counts: dict[str, int]) -> None:
self._counts = counts
self._evaluate_result = False
self._evaluate_values: list[object] = []
class _Locator:
def __init__(self, count_value: int) -> None:
@@ -23,6 +24,8 @@ class _FakePage:
return _FakePage._Locator(self._counts.get(selector, 0))
def evaluate(self, _script: str):
if self._evaluate_values:
return self._evaluate_values.pop(0)
return self._evaluate_result
@@ -40,6 +43,11 @@ class TestListingUnit(unittest.TestCase):
page._evaluate_result = True
self.assertTrue(ListingCollector._has_next_page(page))
def test_get_current_page_number_from_text_counter(self) -> None:
page = _FakePage({})
page._evaluate_values = [5]
self.assertEqual(ListingCollector._get_current_page_number(page), 5)
def test_extract_vehicle_links_from_html_finds_detail_urls(self) -> None:
collector = ListingCollector(Settings(), HumanPacer(Settings()))
html = """

View File

@@ -1,6 +1,7 @@
from __future__ import annotations
import unittest
from types import SimpleNamespace
from unittest.mock import MagicMock
from iaai_scraper.core.config import Settings
@@ -169,6 +170,50 @@ class TestScraperSync(unittest.TestCase):
self.assertTrue(IAAIScraper._is_protection_or_network_error(RuntimeError("captcha challenge")))
self.assertFalse(IAAIScraper._is_protection_or_network_error(RuntimeError("plain validation error")))
def test_recover_empty_listing_page_reload_recovers_links(self) -> None:
scraper = self._make_scraper()
page = MagicMock()
page_result = SimpleNamespace(
vehicle_links=[SimpleNamespace(href="https://www.iaai.com/VehicleDetail/123~US")],
)
scraper.listing_collector.collect_current_page = MagicMock(return_value=page_result)
all_raw_urls: list[str] = []
seen_urls: set[str] = set()
recovered_result, recovered_urls = scraper._recover_empty_listing_page(
page,
page_number=5,
all_raw_urls=all_raw_urls,
seen_urls=seen_urls,
)
page.reload.assert_called_once()
self.assertIs(recovered_result, page_result)
self.assertEqual(recovered_urls, ["https://www.iaai.com/VehicleDetail/123~US"])
self.assertEqual(all_raw_urls, ["https://www.iaai.com/VehicleDetail/123~US"])
def test_recover_empty_listing_page_reopens_listing_for_first_page(self) -> None:
scraper = self._make_scraper()
page = MagicMock()
page_result = SimpleNamespace(
vehicle_links=[SimpleNamespace(href="https://www.iaai.com/VehicleDetail/456~US")],
)
scraper.listing_collector.open_cars_listing = MagicMock()
scraper.listing_collector.collect_current_page = MagicMock(return_value=page_result)
recovered_result, recovered_urls = scraper._recover_empty_listing_page(
page,
page_number=1,
all_raw_urls=[],
seen_urls=set(),
)
scraper.listing_collector.open_cars_listing.assert_called_once_with(page)
page.reload.assert_not_called()
self.assertIs(recovered_result, page_result)
self.assertEqual(recovered_urls, ["https://www.iaai.com/VehicleDetail/456~US"])
if __name__ == "__main__":
unittest.main()

View File

@@ -1,5 +1,6 @@
from __future__ import annotations
import json
import unittest
from unittest.mock import MagicMock, patch
@@ -72,6 +73,7 @@ class TestWorkerTaskLockHelpers(unittest.TestCase):
persistence = MagicMock()
get_persistence.return_value = persistence
redis_client = MagicMock()
redis_client.get.return_value = None
get_redis.return_value = redis_client
stop_event = MagicMock()
heartbeat_thread = MagicMock()
@@ -140,6 +142,7 @@ class TestWorkerTaskLockHelpers(unittest.TestCase):
persistence = MagicMock()
get_persistence.return_value = persistence
redis_client = MagicMock()
redis_client.get.return_value = None
get_redis.return_value = redis_client
stop_event = MagicMock()
heartbeat_thread = MagicMock()
@@ -156,6 +159,78 @@ class TestWorkerTaskLockHelpers(unittest.TestCase):
self.assertEqual(acquire_lock.call_count, 2)
release_lock.assert_called_once()
def test_sync_listing_checkpoint_roundtrip(self) -> None:
redis_client = MagicMock()
storage: dict[str, str] = {}
redis_client.set.side_effect = lambda key, value: storage.__setitem__(key, value)
redis_client.get.side_effect = lambda key: storage.get(key)
tasks._save_sync_checkpoint(
redis_client,
task_id="task-1",
page_number=12,
make=None,
model=None,
lane="iaai_cars",
)
checkpoint = tasks._load_sync_checkpoint(redis_client)
self.assertIsNotNone(checkpoint)
self.assertEqual(checkpoint["last_successful_page"], 12)
self.assertEqual(checkpoint["status"], "in_progress")
def test_sync_listing_task_resumes_from_checkpoint_page(self) -> None:
with patch.object(tasks, "_get_persistence") as get_persistence, \
patch.object(tasks, "_get_redis") as get_redis, \
patch.object(tasks, "_acquire_lock", return_value=True), \
patch.object(tasks, "_is_full_scan_done", return_value=True), \
patch.object(tasks, "_start_lock_heartbeat") as start_heartbeat, \
patch.object(tasks, "_release_lock_if_owner") as release_lock, \
patch.object(tasks, "_clear_sync_checkpoint") as clear_checkpoint, \
patch.object(tasks, "_run_browser_job", side_effect=lambda fn: fn()):
persistence = MagicMock()
get_persistence.return_value = persistence
redis_client = MagicMock()
redis_client.get.side_effect = lambda key: json.dumps({
"status": "in_progress",
"last_successful_page": 9,
"make": None,
"model": None,
"lane": "iaai_cars",
}) if key == tasks.SYNC_LISTING_CHECKPOINT_KEY else None
get_redis.return_value = redis_client
stop_event = MagicMock()
heartbeat_thread = MagicMock()
start_heartbeat.return_value = (stop_event, heartbeat_thread)
sync_listing_mock = MagicMock(return_value={
"run_id": 11,
"status": "success",
"full_scan_completed": True,
"cars_upserted": 1,
"cars_failed": 0,
"images_upserted": 0,
"skipped_existing": 0,
"elapsed_seconds": 1.0,
"failures": [],
})
scraper_ctx = MagicMock()
scraper_ctx.__enter__.return_value.sync_listing = sync_listing_mock
scraper_ctx.__exit__.return_value = None
with patch.object(tasks, "IAAIScraper", return_value=scraper_ctx):
tasks.sync_listing_task.push_request(id="task-789")
try:
result = tasks.sync_listing_task.run()
finally:
tasks.sync_listing_task.pop_request()
self.assertEqual(result["status"], "success")
self.assertEqual(sync_listing_mock.call_args.kwargs["start_page"], 10)
clear_checkpoint.assert_called()
release_lock.assert_called_once()
if __name__ == "__main__":
unittest.main()