Queue gallery enrichment jobs

This commit is contained in:
qananasikq
2026-08-19 12:06:00 +03:00
parent 33b85ccecd
commit 2913199d1e
8 changed files with 487 additions and 212 deletions
+10 -13
View File
@@ -2,13 +2,13 @@ import argparse
from pathlib import Path
from .core.config import Settings
from .core.logs import setup_logging
from .core.utils import save_to_json
from .mobile_de import MobileDeClient, MobileDeScraper
def build_parser() -> argparse.ArgumentParser:
parser = argparse.ArgumentParser(description="mobile.de scraper CLI")
parser.add_argument("--headless", choices=["true", "false"], default=None, help="Override headless mode")
parser.add_argument("--debug", action="store_true", help="Enable DEBUG logging")
subparsers = parser.add_subparsers(dest="command", required=True)
@@ -59,16 +59,13 @@ def main() -> None:
parser = build_parser()
args = parser.parse_args()
runtime_settings: Settings | None = None
if args.headless is not None or args.debug:
runtime_settings = Settings()
if args.headless is not None:
runtime_settings.headless = args.headless == "true"
if args.debug:
runtime_settings.log_level = "DEBUG"
runtime_settings = Settings()
if args.debug:
runtime_settings.log_level = "DEBUG"
setup_logging(runtime_settings.log_level)
if args.command in {"search", "mobilede-search"}:
scraper = MobileDeScraper(MobileDeClient(delay_seconds=args.delay))
scraper = MobileDeScraper(MobileDeClient(delay_seconds=args.delay), runtime_settings=runtime_settings)
data = scraper.collect_search(
start_page=args.page,
max_pages=args.max_pages,
@@ -85,14 +82,14 @@ def main() -> None:
return
if args.command in {"detail", "mobilede-detail"}:
scraper = MobileDeScraper()
scraper = MobileDeScraper(runtime_settings=runtime_settings)
data = scraper.collect_detail(args.listing_id)
save_to_json(data, Path(args.output))
print(f"Saved to {Path(args.output).resolve()}")
return
if args.command == "sync-search":
scraper = MobileDeScraper(MobileDeClient(delay_seconds=args.delay))
scraper = MobileDeScraper(MobileDeClient(delay_seconds=args.delay), runtime_settings=runtime_settings)
data = scraper.sync_search(
start_page=args.page,
max_pages=args.max_pages,
@@ -110,14 +107,14 @@ def main() -> None:
return
if args.command == "sync-detail":
scraper = MobileDeScraper()
scraper = MobileDeScraper(runtime_settings=runtime_settings)
data = scraper.sync_detail(args.listing_id, lane=args.lane)
save_to_json(data, Path(args.output))
print(f"Saved to {Path(args.output).resolve()}")
return
if args.command == "init-db":
scraper = MobileDeScraper()
scraper = MobileDeScraper(runtime_settings=runtime_settings)
data = scraper.init_db()
print(f"DB initialized: {data}")
return