Paper Search MCP

Instructions

Implementation Reference

• paper_search_mcp/server.py:1211-1221 (handler)

  Tool registration for 'read_hal_paper' in server.py.

  @mcp.tool()
  async def read_hal_paper(paper_id: str, save_path: str = "./downloads") -> str:
      """Read and extract text content from a HAL paper.
  
      Args:
          paper_id: HAL paper identifier.
          save_path: Directory where the PDF is/will be saved (default: './downloads').
      Returns:
          str: Extracted text content.
      """
      return hal_searcher.read_paper(paper_id, save_path)

• paper_search_mcp/academic_platforms/hal.py:155-180 (handler)

  Actual implementation logic for reading a HAL paper in HALSearcher class.

  def read_paper(self, paper_id: str, save_path: str = "./downloads") -> str:
      """Download and extract text from a HAL PDF.
  
      Args:
          paper_id: HAL paper ID.
          save_path: Directory where the PDF is/will be saved.
  
      Returns:
          Extracted text content or error message.
      """
      path = self.download_pdf(paper_id, save_path)
      if not path.endswith(".pdf"):
          return path
  
      try:
          try:
              from PyPDF2 import PdfReader
          except ImportError:
              from pypdf import PdfReader
  
          reader = PdfReader(path)
          text_parts = [page.extract_text() for page in reader.pages if page.extract_text()]
          return "\n\n".join(text_parts) if text_parts else "No extractable text in PDF."
      except ImportError:
          return f"PDF downloaded to {path}. Install 'PyPDF2' or 'pypdf' to extract text."
      except Exception as exc: