improve listing sync

iaai_scraper/browser/listing.py

@@ -81,16 +81,20 @@ class ListingCollector:
 
 	def collect_current_page(self, page: Page, page_number: int = 1) -> ListingPageResult:
 		# Считываем ссылки одним проходом по DOM.
-		raw_items = page.eval_on_selector_all(
-			"a[href*='/VehicleDetail/']",
-			"""
-			(nodes) => nodes.map((a) => ({
-				href: a.getAttribute('href') || '',
-				title: a.getAttribute('title') || '',
-				text: (a.textContent || '').trim(),
-			}))
-			""",
-		)
+		try:
+			raw_items = page.eval_on_selector_all(
+				"a[href*='/VehicleDetail/']",
+				"""
+				(nodes) => nodes.map((a) => ({
+					href: a.getAttribute('href') || '',
+					title: a.getAttribute('title') || '',
+					text: (a.textContent || '').trim(),
+				}))
+				""",
+			)
+		except Exception as exc:
+			logger.warning("collect_current_page failed on page %d: %s", page_number, exc)
+			raw_items = []
 		total = min(len(raw_items), self.settings.listing.page_link_limit)
 		links: list[ListingVehicleLink] = []
 		seen: set[str] = set()
@@ -127,13 +131,19 @@ class ListingCollector:
 			locator = page.locator(selector).first
 			if locator.count() == 0:
 				continue
-			disabled = (locator.get_attribute("disabled") or "").lower()
-			aria_disabled = (locator.get_attribute("aria-disabled") or "").lower()
-			classes = (locator.get_attribute("class") or "").lower()
+			try:
+				disabled = (locator.get_attribute("disabled", timeout=1500) or "").lower()
+				aria_disabled = (locator.get_attribute("aria-disabled", timeout=1500) or "").lower()
+				classes = (locator.get_attribute("class", timeout=1500) or "").lower()
+			except Exception:
+				continue
 			if disabled or aria_disabled == "true" or "disabled" in classes:
 				continue
-			self.pacer.move_mouse_to(page, locator)
-			locator.click()
+			try:
+				self.pacer.move_mouse_to(page, locator)
+				locator.click(timeout=8000)
+			except Exception:
+				continue
 
 			# Ждём смены контента (AJAX пагинация): первая VehicleDetail-ссылка должна измениться.
 			if old_first_href:
@@ -168,15 +178,29 @@ class ListingCollector:
 		make: str | None = None,
 		model: str | None = None,
 		known_origin_ids: set[str] | None = None,
+		max_duration_seconds: float | None = None,
 	) -> dict[str, Any]:
 		self.open_cars_listing(page)
 		applied_filters = self.apply_filters(page, make=make, model=model)
+		started_at = time.perf_counter()
+		truncated_by_time_budget = False
 		pages: list[dict[str, object]] = []
 		all_links: list[str] = []
 		early_stopped = False
 		threshold = self.settings.listing.early_stop_threshold
 
 		for page_number in range(1, max(1, self.settings.listing.max_pages_per_run) + 1):
+			if max_duration_seconds is not None and max_duration_seconds > 0:
+				elapsed = time.perf_counter() - started_at
+				if elapsed >= max_duration_seconds:
+					truncated_by_time_budget = True
+					logger.warning(
+						"Listing collection stopped by time budget: page=%d elapsed=%.1fs budget=%.1fs",
+						page_number,
+						elapsed,
+						max_duration_seconds,
+					)
+					break
 			page_result = self.collect_current_page(page, page_number=page_number)
 			pages.append({
 				"page_number": page_result.page_number,
@@ -236,6 +260,7 @@ class ListingCollector:
 			"vehicles_collected": len(all_links),
 			"vehicle_urls": all_links,
 			"early_stopped": early_stopped,
+			"truncated_by_time_budget": truncated_by_time_budget,
 			"pages": pages,
 			"strategy": {
 				"sequential": True,