This commit is contained in:
lasta
2026-05-22 00:22:03 +09:00
parent 8d77bc659f
commit d841fb823a
49 changed files with 2732 additions and 3763 deletions

View File

@@ -37,6 +37,8 @@ def build_parser() -> argparse.ArgumentParser:
crawl.add_argument("--extractor-provider", default="rule_based", choices=["rule_based", "openai", "ollama", "lm_studio"])
crawl.add_argument("--extractor-model")
crawl.add_argument("--extractor-base-url")
crawl.add_argument("--extraction-mode", default=None, choices=["rule_only", "llm_only", "hybrid", "compare"])
crawl.add_argument("--no-rule-fallback", action="store_true")
claims = sub.add_parser("claims")
claims.add_argument("--project", required=True)
@@ -81,6 +83,8 @@ def main() -> None:
provider=args.extractor_provider,
model=args.extractor_model,
base_url=args.extractor_base_url,
extraction_mode=args.extraction_mode,
fallback_to_rules=not args.no_rule_fallback,
),
)
result = pipeline.crawl_url(config, args.source, args.url)