Paper Search MCP

Instructions

Implementation Reference

• paper_search_mcp/academic_platforms/citeseerx.py:344-365 (handler)

  The actual implementation of the read_paper logic for CiteSeerX.

  def read_paper(self, paper_id: str, save_path: str = "./downloads") -> str:
      """
      Download and extract text from a CiteSeerX paper.
  
      Note: CiteSeerX provides abstracts but not always full text.
      This method tries to download PDF and extract text if available.
  
      Args:
          paper_id: CiteSeerX paper identifier
          save_path: Directory where PDF is/will be saved
  
      Returns:
          Extracted text content of the paper (abstract if PDF not available)
  
      Raises:
          Exception: If paper reading fails
      """
      try:
          # First get paper details
          paper = self.get_paper_details(paper_id)
          if not paper:
              raise Exception(f"Paper {paper_id} not found")

• paper_search_mcp/server.py:1107-1117 (registration)

  Registration and tool definition for read_citeseerx_paper in the MCP server.

  @mcp.tool()
  async def read_citeseerx_paper(paper_id: str, save_path: str = "./downloads") -> str:
      """Read and extract text content from a CiteSeerX paper.
  
      Args:
          paper_id: CiteSeerX paper identifier.
          save_path: Directory where the PDF is/will be saved (default: './downloads').
      Returns:
          str: Extracted text or fallback abstract/error message.
      """
      return citeseerx_searcher.read_paper(paper_id, save_path)