bc_search_google_scholar_publications

Instructions

Implementation Reference

• src/biocontext_kb/core/scholarly/_search_publications.py:12-71 (handler)

  The actual tool handler function 'search_google_scholar_publications' decorated with @core_mcp.tool(). Searches Google Scholar for publications using the scholarly library, supports proxy configuration, author search syntax, and returns publication metadata (title, author, venue, year, citations, abstract, URLs).

  @core_mcp.tool()
  def search_google_scholar_publications(
      query: Annotated[
          str,
          Field(description="Search query (e.g., 'machine learning' or 'author:\"John Smith\" deep learning')"),
      ],
      max_results: Annotated[int, Field(description="Maximum number of publications to return (1-50)", ge=1, le=50)] = 10,
      use_proxy: Annotated[bool, Field(description="Use free proxies to avoid rate limiting")] = True,
  ) -> Dict[str, Any]:
      """Search Google Scholar for publications with support for author search using 'author:"Name"' syntax. WARNING: Use responsibly, may block excessive queries.
  
      Returns:
          dict: Publications list with title, authors, venue, year, citations, abstract, bib entry or error message.
      """
      try:
          # Set up proxy if requested
          if use_proxy:
              try:
                  pg = ProxyGenerator()
                  pg.FreeProxies()
                  scholarly.use_proxy(pg)
                  logger.info("Proxy configured for Google Scholar requests")
              except Exception as e:
                  logger.warning(f"Failed to set up proxy: {e}")
                  # Continue without proxy
  
          # Search for publications
          search_query = scholarly.search_pubs(query)
  
          publications = []
  
          for count, pub in enumerate(search_query):
              if count >= max_results:
                  break
  
              # Extract publication information
              bib = pub.get("bib", {})
              pub_info = {
                  "title": bib.get("title", ""),
                  "author": bib.get("author", ""),
                  "venue": bib.get("venue", ""),
                  "pub_year": bib.get("pub_year", ""),
                  "abstract": bib.get("abstract", ""),
                  "pub_url": bib.get("pub_url", ""),
                  "eprint_url": pub.get("eprint_url", ""),
                  "num_citations": pub.get("num_citations", 0),
                  "citedby_url": pub.get("citedby_url", ""),
                  "url_scholarbib": pub.get("url_scholarbib", ""),
              }
  
              publications.append(pub_info)
  
          return {"query": query, "total_found": len(publications), "publications": publications}
  
      except Exception as e:
          logger.error(f"Error searching Google Scholar publications: {e}")
          return {
              "error": f"Failed to search Google Scholar publications: {e!s}",
              "note": "Google Scholar may be blocking requests. Publication searches are particularly risky. Try again later or use alternative databases like PubMed/EuropePMC.",
          }

• src/biocontext_kb/core/scholarly/_search_publications.py:13-19 (schema)

  Input schema for the tool: query (str, required), max_results (int, 1-50, default 10), use_proxy (bool, default True) — all defined using Pydantic Field annotations.

  def search_google_scholar_publications(
      query: Annotated[
          str,
          Field(description="Search query (e.g., 'machine learning' or 'author:\"John Smith\" deep learning')"),
      ],
      max_results: Annotated[int, Field(description="Maximum number of publications to return (1-50)", ge=1, le=50)] = 10,
      use_proxy: Annotated[bool, Field(description="Use free proxies to avoid rate limiting")] = True,

• src/biocontext_kb/core/scholarly/_search_publications.py:12-12 (registration)

  Tool registration via @core_mcp.tool() decorator on the function. The 'core_mcp' FastMCP server instance is imported from biocontext_kb/core/_server.py.

  @core_mcp.tool()

• src/biocontext_kb/core/_server.py:3-6 (registration)

  The FastMCP server instance 'core_mcp' that registers the tool via the @core_mcp.tool() decorator.

  core_mcp = FastMCP(  # type: ignore
      "BC",
      instructions="Provides access to biomedical knowledge bases.",
  )

• src/biocontext_kb/core/scholarly/__init__.py:9-13 (helper)

  Re-exports search_google_scholar_publications from the _search_publications module, making it part of the scholarly package's public API.

  from ._search_publications import search_google_scholar_publications
  
  __all__ = [
      "search_google_scholar_publications",
  ]