Paper Search MCP

Instructions

Implementation Reference

• paper_search_mcp/server.py:494-504 (handler)

  MCP tool definition and handler for download_medrxiv in the server.

  @mcp.tool()
  async def download_medrxiv(paper_id: str, save_path: str = "./downloads") -> str:
      """Download PDF of a medRxiv paper.
  
      Args:
          paper_id: medRxiv DOI.
          save_path: Directory to save the PDF (default: './downloads').
      Returns:
          Path to the downloaded PDF file.
      """
      return medrxiv_searcher.download_pdf(paper_id, save_path)

• paper_search_mcp/academic_platforms/medrxiv.py:87-115 (handler)

  The actual implementation of the medRxiv PDF download logic within MedRxivSearcher.

  def download_pdf(self, paper_id: str, save_path: str) -> str:
      """
      Download a PDF for a given paper ID from medRxiv.
  
      Args:
          paper_id: The DOI of the paper.
          save_path: Directory to save the PDF.
  
      Returns:
          Path to the downloaded PDF file.
      """
      if not paper_id:
          raise ValueError("Invalid paper_id: paper_id is empty")
  
      pdf_url = f"https://www.medrxiv.org/content/{paper_id}v1.full.pdf"
      tries = 0
      while tries < self.max_retries:
          try:
              # Add User-Agent to avoid potential 403 errors
              headers = {
                  'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/91.0.4472.124 Safari/537.36'
              }
              response = self.session.get(pdf_url, timeout=self.timeout, headers=headers)
              response.raise_for_status()
              os.makedirs(save_path, exist_ok=True)
              output_file = f"{save_path}/{paper_id.replace('/', '_')}.pdf"
              with open(output_file, 'wb') as f:
                  f.write(response.content)
              return output_file