Files
encar/encar_scraper/cli.py
2026-04-16 18:08:31 +03:00

104 lines
4.6 KiB
Python

import argparse
from pathlib import Path
from .core.config import Settings
from .core.utils import save_to_json
from .encar import EncarScraper, EncarFilters, CAR_TYPE_ALL, CAR_TYPE_DOMESTIC, CAR_TYPE_IMPORT
def build_parser() -> argparse.ArgumentParser:
parser = argparse.ArgumentParser(description="Encar scraper CLI")
subparsers = parser.add_subparsers(dest="command", required=True)
default_output_dir = Path("artifacts/json")
subparsers.add_parser("init-db", help="Create DB tables")
# collect-listing
collect_parser = subparsers.add_parser("collect-listing", help="Collect Encar listing")
collect_parser.add_argument("--limit", type=int, default=None, help="Max items to collect")
collect_parser.add_argument("--car-type", choices=["all", "domestic", "import"], default="all",
help="all=все, domestic=корейские, import=импорт")
collect_parser.add_argument("--manufacturer", default=None, help="Filter by manufacturer (e.g. BMW, 현대)")
collect_parser.add_argument("--year-from", type=int, default=None, help="Min year")
collect_parser.add_argument("--year-to", type=int, default=None, help="Max year")
collect_parser.add_argument("--price-max", type=int, default=None, help="Max price in 만원 (10k KRW)")
collect_parser.add_argument("--output", default=str(default_output_dir / "encar_listing.json"))
# scrape-vehicle
scrape_parser = subparsers.add_parser("scrape-vehicle", help="Fetch single vehicle data")
scrape_parser.add_argument("vehicle_url")
scrape_parser.add_argument("--output", default=str(default_output_dir / "encar_vehicle_detail.json"))
# sync-vehicle
sync_parser = subparsers.add_parser("sync-vehicle", help="Sync single vehicle into DB")
sync_parser.add_argument("vehicle_url")
sync_parser.add_argument("--lane", default="encar")
sync_parser.add_argument("--output", default=str(default_output_dir / "encar_sync_vehicle.json"))
# sync-listing
sync_listing_parser = subparsers.add_parser("sync-listing", help="Collect listing and sync all to DB")
sync_listing_parser.add_argument("--limit", type=int, default=None, help="Max items to sync")
sync_listing_parser.add_argument("--car-type", choices=["all", "domestic", "import"], default="all")
sync_listing_parser.add_argument("--manufacturer", default=None, help="Filter by manufacturer")
sync_listing_parser.add_argument("--year-from", type=int, default=None)
sync_listing_parser.add_argument("--year-to", type=int, default=None)
sync_listing_parser.add_argument("--price-max", type=int, default=None)
sync_listing_parser.add_argument("--lane", default="encar")
sync_listing_parser.add_argument("--only-new", choices=["true", "false"], default="true")
sync_listing_parser.add_argument("--output", default=str(default_output_dir / "encar_sync_listing.json"))
return parser
def _build_filters_from_args(args) -> EncarFilters:
"""Строит фильтры из аргументов CLI."""
car_type_map = {"all": CAR_TYPE_ALL, "domestic": CAR_TYPE_DOMESTIC, "import": CAR_TYPE_IMPORT}
return EncarFilters(
car_type=car_type_map.get(getattr(args, "car_type", "all"), CAR_TYPE_ALL),
manufacturer=getattr(args, "manufacturer", None),
year_from=getattr(args, "year_from", None),
year_to=getattr(args, "year_to", None),
price_to=getattr(args, "price_max", None),
)
def main() -> None:
import sys
import logging
logging.basicConfig(
level=logging.INFO,
format="%(asctime)s [%(levelname)s] %(name)s: %(message)s",
stream=sys.stderr,
)
parser = build_parser()
args = parser.parse_args()
scraper = EncarScraper()
if args.command == "init-db":
result = scraper.init_db()
print(f"DB initialized: {result}")
return
elif args.command == "collect-listing":
filters = _build_filters_from_args(args)
result = scraper.collect_listing(limit=args.limit, filters=filters)
elif args.command == "scrape-vehicle":
result = scraper.scrape_vehicle_detail(args.vehicle_url)
elif args.command == "sync-vehicle":
result = scraper.sync_vehicle(args.vehicle_url, lane=args.lane)
elif args.command == "sync-listing":
filters = _build_filters_from_args(args)
only_new = args.only_new == "true"
result = scraper.sync_listing(limit=args.limit, filters=filters, lane=args.lane, only_new=only_new)
else:
raise SystemExit(f"Unknown command: {args.command}")
save_to_json(result, Path(args.output))
print(f"Saved to {Path(args.output).resolve()}")
if __name__ == "__main__":
main()