sync only new cars

.env.example

@@ -36,6 +36,7 @@ IAAI_AFTER_PAGE_CHANGE_MAX_S=6.0
 
 # Scheduler (default once per hour)
 IAAI_SCHEDULER_INTERVAL_MINUTES=60
+IAAI_SYNC_ONLY_NEW=true
 
 # Retry / backoff
 IAAI_RETRY_DELAY_SECONDS=2.5

iaai_scraper/cli.py

@@ -13,39 +13,41 @@ def build_parser() -> argparse.ArgumentParser:
     parser.add_argument("--debug", action="store_true", help="Enable DEBUG logging")
     subparsers = parser.add_subparsers(dest="command", required=True)
 
+    default_output_dir = Path("artifacts/json")
+
     init_db_parser = subparsers.add_parser("init-db", help="Create local DB tables")
-    init_db_parser.add_argument("--output", default="iaai_db_init.json", help="Path to output JSON")
+    init_db_parser.add_argument("--output", default=str(default_output_dir / "iaai_db_init.json"), help="Path to output JSON")
 
     listing_parser = subparsers.add_parser("collect-listing", help="Collect vehicle URLs from Vehiclelisting/Cars")
     listing_parser.add_argument("--make", default=None, help="Optional make filter")
     listing_parser.add_argument("--model", default=None, help="Optional model filter")
-    listing_parser.add_argument("--output", default="iaai_listing_links.json", help="Path to output JSON")
+    listing_parser.add_argument("--output", default=str(default_output_dir / "iaai_listing_links.json"), help="Path to output JSON")
 
     open_parser = subparsers.add_parser(
         "open-vehicle",
         help="Open a vehicle page in gentle mode and save only DOM-based hints",
     )
     open_parser.add_argument("vehicle_url", help="IAAI vehicle detail URL")
-    open_parser.add_argument("--output", default="iaai_vehicle_opened.json", help="Path to output JSON")
+    open_parser.add_argument("--output", default=str(default_output_dir / "iaai_vehicle_opened.json"), help="Path to output JSON")
 
     scrape_parser = subparsers.add_parser(
         "scrape-vehicle",
         help="Open a vehicle page and capture a limited set of likely useful JSON responses",
     )
     scrape_parser.add_argument("vehicle_url", help="IAAI vehicle detail URL")
-    scrape_parser.add_argument("--output", default="iaai_vehicle_detail.json", help="Path to output JSON")
+    scrape_parser.add_argument("--output", default=str(default_output_dir / "iaai_vehicle_detail.json"), help="Path to output JSON")
 
     export_parser = subparsers.add_parser(
         "export-db-json",
         help="Scrape a vehicle page and save only the DB-ready car record JSON",
     )
     export_parser.add_argument("vehicle_url", help="IAAI vehicle detail URL")
-    export_parser.add_argument("--output", default="iaai_vehicle_db_record.json", help="Path to output JSON")
+    export_parser.add_argument("--output", default=str(default_output_dir / "iaai_vehicle_db_record.json"), help="Path to output JSON")
 
     sync_vehicle_parser = subparsers.add_parser("sync-vehicle", help="Scrape one vehicle and upsert it into the DB")
     sync_vehicle_parser.add_argument("vehicle_url", help="IAAI vehicle detail URL")
     sync_vehicle_parser.add_argument("--lane", default="iaai", help="Logical lane name for sync_runs")
-    sync_vehicle_parser.add_argument("--output", default="iaai_sync_vehicle.json", help="Path to output JSON")
+    sync_vehicle_parser.add_argument("--output", default=str(default_output_dir / "iaai_sync_vehicle.json"), help="Path to output JSON")
 
     sync_listing_parser = subparsers.add_parser(
         "sync-listing",
@@ -55,7 +57,13 @@ def build_parser() -> argparse.ArgumentParser:
     sync_listing_parser.add_argument("--model", default=None, help="Optional model filter")
     sync_listing_parser.add_argument("--lane", default="iaai_cars", help="Logical lane name for sync_runs")
     sync_listing_parser.add_argument("--limit", type=int, default=None, help="Limit number of vehicles to sync")
-    sync_listing_parser.add_argument("--output", default="iaai_sync_listing.json", help="Path to output JSON")
+    sync_listing_parser.add_argument(
+        "--only-new",
+        choices=["true", "false"],
+        default=None,
+        help="Process only new vehicles (default from IAAI_SYNC_ONLY_NEW)",
+    )
+    sync_listing_parser.add_argument("--output", default=str(default_output_dir / "iaai_sync_listing.json"), help="Path to output JSON")
 
     daemon_parser = subparsers.add_parser(
         "run-daemon",
@@ -108,7 +116,14 @@ def main() -> None:
         elif args.command == "sync-vehicle":
             data = scraper.sync_vehicle(args.vehicle_url, lane=args.lane)
         else:
-            data = scraper.sync_listing(make=args.make, model=args.model, lane=args.lane, limit=args.limit)
+            only_new = None if args.only_new is None else args.only_new == "true"
+            data = scraper.sync_listing(
+                make=args.make,
+                model=args.model,
+                lane=args.lane,
+                limit=args.limit,
+                only_new=only_new,
+            )
 
         save_to_json(data, Path(args.output))
         print(f"Saved result to {Path(args.output).resolve()}")

iaai_scraper/core/config.py

@@ -126,6 +126,7 @@ class Settings:
     log_file: str | None = os.getenv("IAAI_LOG_FILE") or None
     enable_trace_id_logs: bool = os.getenv("IAAI_ENABLE_TRACE_ID_LOGS", "true").strip().lower() in {"1", "true", "yes", "on"}
     scheduler_interval_minutes: int = int(os.getenv("IAAI_SCHEDULER_INTERVAL_MINUTES", "60"))
+    sync_only_new: bool = os.getenv("IAAI_SYNC_ONLY_NEW", "true").strip().lower() in {"1", "true", "yes", "on"}
     fingerprint: FingerprintConfig = field(default_factory=FingerprintConfig)
     gentle: GentleModeConfig = field(default_factory=GentleModeConfig)
     pace: HumanPaceConfig = field(default_factory=HumanPaceConfig)

iaai_scraper/scraper.py

@@ -1,4 +1,5 @@
 import logging
+import re
 import signal
 import time
 import uuid
@@ -21,10 +22,18 @@ from .storage.listing import ListingCollector
 from .storage.schemas import CarRecord
 
 logger = logging.getLogger("iaai_scraper.scraper")
+VEHICLE_ID_RE = re.compile(r"/VehicleDetail/(\d+)(?:~[A-Z]{2})?", re.IGNORECASE)
 
 
 class IAAIScraper:
 
+    @staticmethod
+    def _extract_origin_id_from_url(vehicle_url: str) -> str | None:
+        match = VEHICLE_ID_RE.search(vehicle_url)
+        if not match:
+            return None
+        return match.group(1)
+
     def __init__(self, runtime_settings: Settings | None = None) -> None:
         # Базовые зависимости и сервисы скрапера.
         self.settings = runtime_settings or settings
@@ -260,7 +269,14 @@ class IAAIScraper:
                 error_summary=error_summary,
             )
 
-    def sync_listing(self, make: str | None = None, model: str | None = None, lane: str = "iaai_cars", limit: int | None = None):
+    def sync_listing(
+        self,
+        make: str | None = None,
+        model: str | None = None,
+        lane: str = "iaai_cars",
+        limit: int | None = None,
+        only_new: bool | None = None,
+    ):
         """Листинг + sync всех найденных машин."""
         # Массовая синхронизация с общим run_id и сбором ошибок.
         trace_id = self._new_trace_id("sync-listing")
@@ -271,6 +287,7 @@ class IAAIScraper:
         cars_failed = 0
         images_upserted = 0
         total = 0
+        skipped_existing = 0
         failures: list[dict[str, str]] = []
         listing: dict = {}
 
@@ -280,6 +297,27 @@ class IAAIScraper:
             if limit is not None:
                 vehicle_urls = vehicle_urls[:max(0, limit)]
 
+            effective_only_new = self.settings.sync_only_new if only_new is None else only_new
+            if effective_only_new:
+                existing_urls = self.persistence.get_existing_origin_urls(vehicle_urls)
+                url_to_origin_id = {
+                    url: self._extract_origin_id_from_url(url)
+                    for url in vehicle_urls
+                }
+                candidate_origin_ids = [origin_id for origin_id in url_to_origin_id.values() if origin_id]
+                existing_ids = self.persistence.get_existing_origin_ids(candidate_origin_ids)
+
+                known_urls = {
+                    url
+                    for url in vehicle_urls
+                    if (url in existing_urls) or (url_to_origin_id.get(url) in existing_ids)
+                }
+
+                skipped_existing = len(known_urls)
+                if skipped_existing:
+                    logger.info("Filtering already known vehicles: skipped %d", skipped_existing)
+                vehicle_urls = [url for url in vehicle_urls if url not in known_urls]
+
             total = len(vehicle_urls)
             logger.info("Starting sync: %d vehicles to process", total)
 
@@ -345,6 +383,7 @@ class IAAIScraper:
             "cars_upserted": cars_upserted,
             "cars_failed": cars_failed,
             "images_upserted": images_upserted,
+            "skipped_existing": skipped_existing,
             "elapsed_seconds": round(time.perf_counter() - started_at, 3),
             "failures": failures,
         }

iaai_scraper/storage/db.py

@@ -4,7 +4,7 @@ from contextlib import contextmanager
 from datetime import datetime, timezone
 from typing import Iterator
 
-from sqlalchemy import create_engine, select
+from sqlalchemy import create_engine, or_, select
 from sqlalchemy.orm import Session, sessionmaker
 
 from ..core.config import Settings
@@ -76,6 +76,16 @@ class PersistenceService:
 	def start_sync_run(self, lane: str) -> int:
 		# Создаём запись о запуске синхронизации.
 		with self.session_scope() as session:
+			# Если предыдущий процесс умер, оставив run в `running`,
+			# помечаем его как failed перед новым запуском.
+			now = datetime.now(timezone.utc)
+			stale_runs = session.execute(select(SyncRun).where(SyncRun.status == "running")).scalars().all()
+			for stale in stale_runs:
+				stale.status = "failed"
+				stale.finished_at = now
+				if not stale.error_summary:
+					stale.error_summary = "Recovered stale running sync run before starting a new run"
+
 			run = SyncRun(status="running", lane=lane, ids_fetched=0, cars_upserted=0, cars_failed=0, images_upserted=0)
 			session.add(run)
 			session.flush()
@@ -95,6 +105,22 @@ class PersistenceService:
 			run.images_upserted = images_upserted
 			run.error_summary = error_summary
 
+	def get_existing_origin_urls(self, origin_urls: list[str]) -> set[str]:
+		# Возвращает уже существующие в БД origin_url для фильтрации only-new запусков.
+		if not origin_urls:
+			return set()
+		with self.session_scope() as session:
+			rows = session.execute(select(Car.origin_url).where(Car.origin_url.in_(origin_urls))).all()
+			return {str(row[0]) for row in rows if row and row[0]}
+
+	def get_existing_origin_ids(self, origin_ids: list[str]) -> set[str]:
+		# Возвращает уже существующие в БД origin_id для фильтрации только новых авто.
+		if not origin_ids:
+			return set()
+		with self.session_scope() as session:
+			rows = session.execute(select(Car.origin_id).where(Car.origin_id.in_(origin_ids))).all()
+			return {str(row[0]) for row in rows if row and row[0]}
+
 	@staticmethod
 	def _add_images(session: Session, car_id: int, images: list[dict[str, object]]) -> None:
 		for image_payload in images:
@@ -104,7 +130,7 @@ class PersistenceService:
 	def _car_payload(record: CarRecord) -> dict[str, object]:
 		payload = record.model_dump(mode="python")
 		result = {key: value for key, value in payload.items() if key in CAR_DB_FIELDS}
-		# Serialize raw_attributes dict to JSON string for Text column.
+	
 		if "raw_attributes" in result and isinstance(result["raw_attributes"], dict):
 			result["raw_attributes"] = json.dumps(result["raw_attributes"], ensure_ascii=False, default=str)
 		return result
@@ -116,8 +142,11 @@ class PersistenceService:
 		images = [image.model_dump(mode="python") for image in record.images]
 		content_hash = str(payload.get("content_hash") or "")
 		with self.session_scope() as session:
-			# поиск по origin_id
+			# Сначала пытаемся найти по origin_id, а если ранее origin_id был неполный,
-			car = session.execute(select(Car).where(Car.origin_id == record.origin_id)).scalar_one_or_none()
+			# подхватываем существующую запись по origin_url, чтобы не плодить дубли.
+			car = session.execute(
+				select(Car).where(or_(Car.origin_id == record.origin_id, Car.origin_url == record.origin_url))
+			).scalar_one_or_none()
 			action = "inserted"
 			if car is None:
 				car = Car(**payload)