firecrawl.py 1.5 KB

12345678910111213141516171819202122232425262728293031323334353637383940414243444546474849
  1. import logging
  2. from typing import Optional, List
  3. from urllib.parse import urljoin
  4. import requests
  5. from open_webui.retrieval.web.main import SearchResult, get_filtered_results
  6. from open_webui.env import SRC_LOG_LEVELS
  7. log = logging.getLogger(__name__)
  8. log.setLevel(SRC_LOG_LEVELS["RAG"])
  9. def search_firecrawl(
  10. firecrawl_url: str,
  11. firecrawl_api_key: str,
  12. query: str,
  13. count: int,
  14. filter_list: Optional[List[str]] = None,
  15. ) -> List[SearchResult]:
  16. try:
  17. firecrawl_search_url = urljoin(firecrawl_url, "/v1/search")
  18. response = requests.post(
  19. firecrawl_search_url,
  20. headers={
  21. "User-Agent": "Open WebUI (https://github.com/open-webui/open-webui) RAG Bot",
  22. "Authorization": f"Bearer {firecrawl_api_key}",
  23. },
  24. json={
  25. "query": query,
  26. "limit": count,
  27. },
  28. )
  29. response.raise_for_status()
  30. results = response.json().get("data", [])
  31. if filter_list:
  32. results = get_filtered_results(results, filter_list)
  33. results = [
  34. SearchResult(
  35. link=result.get("url"),
  36. title=result.get("title"),
  37. snippet=result.get("description"),
  38. )
  39. for result in results[:count]
  40. ]
  41. log.info(f"External search results: {results}")
  42. return results
  43. except Exception as e:
  44. log.error(f"Error in External search: {e}")
  45. return []