Paper Search MCP Server

Instructions

DOWNLOAD PRIORITY (try in order):
1. If arXiv paper -> use read_arxiv_paper(arxiv_id)
2. If published before 2023 -> use read_scihub_paper(doi)
3. Use this tool as last resort

Args:
    paper_id: Semantic Scholar ID or prefixed ID (DOI:, ARXIV:, PMID:).
    save_path: Directory to save PDF.

Returns:
    Full paper text in Markdown format.

Implementation Reference

• paper_find_mcp/server.py:602-618 (handler)

  MCP tool handler for 'read_semantic_paper'. This is the primary entry point decorated with @mcp.tool(). It delegates execution to the generic _read helper using the 'semantic' searcher instance.

  @mcp.tool()
  async def read_semantic_paper(paper_id: str, save_path: Optional[str] = None) -> str:
      """Read paper via Semantic Scholar (open-access only, use as LAST RESORT).
      
      DOWNLOAD PRIORITY (try in order):
      1. If arXiv paper -> use read_arxiv_paper(arxiv_id)
      2. If published before 2023 -> use read_scihub_paper(doi)
      3. Use this tool as last resort
      
      Args:
          paper_id: Semantic Scholar ID or prefixed ID (DOI:, ARXIV:, PMID:).
          save_path: Directory to save PDF.
      
      Returns:
          Full paper text in Markdown format.
      """
      return await _read('semantic', paper_id, save_path)

• paper_find_mcp/server.py:137-157 (helper)

  Generic helper function _read that retrieves the searcher instance from SEARCHERS dict and invokes searcher.read_paper(paper_id, save_path). Used by all platform-specific read_*_paper tools.

  async def _read(
      searcher_name: str, 
      paper_id: str, 
      save_path: Optional[str] = None
  ) -> str:
      """通用阅读函数"""
      if save_path is None:
          save_path = get_download_path()
      
      searcher = SEARCHERS.get(searcher_name)
      if not searcher:
          return f"Error: Unknown searcher {searcher_name}"
      
      try:
          return searcher.read_paper(paper_id, save_path)
      except NotImplementedError as e:
          return str(e)
      except Exception as e:
          logger.error(f"Read failed for {searcher_name}: {e}")
          return f"Error reading paper: {str(e)}"

• paper_find_mcp/academic_platforms/semantic.py:379-421 (helper)

  Core implementation logic in SemanticSearcher.read_paper(). Downloads open-access PDF using get_paper_details() and pdf_url, then extracts full text to Markdown using pymupdf4llm.to_markdown(), prepends metadata.

  def read_paper(self, paper_id: str, save_path: str) -> str:
      """下载并提取论文文本
      
      使用 PyMuPDF4LLM 提取 Markdown 格式。
      
      Args:
          paper_id: 论文 ID
          save_path: 保存目录
          
      Returns:
          提取的文本内容或错误信息
      """
      # 先下载 PDF
      pdf_path = self.download_pdf(paper_id, save_path)
      if pdf_path.startswith("Error"):
          return pdf_path
      
      # 获取论文元数据
      paper = self.get_paper_details(paper_id)
      
      try:
          text = pymupdf4llm.to_markdown(pdf_path, show_progress=False)
          logger.info(f"Extracted {len(text)} characters using PyMuPDF4LLM")
          
          if not text.strip():
              return f"PDF downloaded to {pdf_path}, but no text could be extracted."
          
          # 添加元数据
          metadata = ""
          if paper:
              metadata = f"# {paper.title}\n\n"
              metadata += f"**Authors**: {', '.join(paper.authors)}\n"
              metadata += f"**Published**: {paper.published_date}\n"
              metadata += f"**URL**: {paper.url}\n"
              metadata += f"**PDF**: {pdf_path}\n\n"
              metadata += "---\n\n"
          
          return metadata + text
          
      except Exception as e:
          logger.error(f"Failed to extract text: {e}")
          return f"Error extracting text: {e}"

• paper_find_mcp/server.py:75-85 (registration)

  Global SEARCHERS dictionary where 'semantic': SemanticSearcher() instance is registered, used by _read and _download helpers to dispatch to platform-specific implementations.

  SEARCHERS = {
      'arxiv': ArxivSearcher(),
      'pubmed': PubMedSearcher(),
      'biorxiv': BioRxivSearcher(),
      'medrxiv': MedRxivSearcher(),
      'google_scholar': GoogleScholarSearcher(),
      'iacr': IACRSearcher(),
      'semantic': SemanticSearcher(),
      'crossref': CrossRefSearcher(),
      'repec': RePECSearcher(),
  }