Paper Search MCP

Instructions

Implementation Reference

• paper_search_mcp/server.py:1199-1208 (handler)

  The MCP tool handler function `download_zenodo` in the server.py, which calls the ZenodoSearcher helper.

  async def download_zenodo(paper_id: str, save_path: str = "./downloads") -> str:
      """Download PDF for a paper from Zenodo.
  
      Args:
          paper_id: Zenodo paper identifier.
          save_path: Directory to save the PDF (default: './downloads').
      Returns:
          str: Path to downloaded PDF.
      """
      return zenodo_searcher.download_pdf(paper_id, save_path)

• paper_search_mcp/academic_platforms/zenodo.py:113-158 (handler)

  The actual implementation of `download_pdf` within the `ZenodoSearcher` class, which handles the file downloading logic.

  def download_pdf(self, paper_id: str, save_path: str = "./downloads") -> str:
      """Download an open-access PDF from Zenodo.
  
      Args:
          paper_id: Zenodo record ID (numeric string or full DOI ``10.5281/zenodo.NNNNNN``).
          save_path: Directory to save the PDF.
  
      Returns:
          Absolute path to the saved PDF, or an error message.
      """
      import re
      import os
  
      record_id = self._extract_record_id(paper_id)
      if not record_id:
          return f"Could not determine Zenodo record ID from: {paper_id}"
  
      try:
          response = self.session.get(
              f"{self.BASE_URL}/records/{record_id}", timeout=20
          )
          response.raise_for_status()
          record = response.json()
      except Exception as exc:
          return f"Failed to fetch Zenodo record {record_id}: {exc}"
  
      pdf_url = self._find_pdf_url(record)
      if not pdf_url:
          return (
              f"No open-access PDF found for Zenodo record {record_id}.  "
              "The record may be embargoed or restricted."
          )
  
      os.makedirs(save_path, exist_ok=True)
      safe_name = re.sub(r"[^a-zA-Z0-9._-]+", "_", record_id) or record_id
      output_path = os.path.join(save_path, f"zenodo_{safe_name}.pdf")
  
      try:
          dl_response = self.session.get(pdf_url, stream=True, timeout=60)
          dl_response.raise_for_status()
          with open(output_path, "wb") as fh:
              for chunk in dl_response.iter_content(chunk_size=8192):
                  fh.write(chunk)
          return output_path
      except Exception as exc:
          return f"Failed to download PDF from {pdf_url}: {exc}"