Huggingface Daily Papers

Instructions

Implementation Reference

• main.py:133-161 (handler)

  MCP server.call_tool handler implementation for 'get_today_papers': fetches papers from scraper, handles empty case, formats detailed markdown list of papers with title, authors, abstract, URL, PDF, votes, and submitter.

  elif name == "get_today_papers":
      papers = scraper.get_today_papers()
      today = datetime.now().strftime("%Y-%m-%d")
      
      if not papers:
          return [
              types.TextContent(
                  type="text",
                  text=f"No papers found for today ({today}). Papers might not be published yet or there could be a network issue."
              )
          ]
      
      return [
          types.TextContent(
              type="text", 
              text=f"Today's Papers ({today}) - Found {len(papers)} papers:\n\n" +
                   "\n".join([
                       f"Title: {paper['title']}\n"
                       f"Authors: {', '.join(paper['authors'])}\n"
                       f"Abstract: {paper['abstract']}\n" 
                       f"URL: {paper['url']}\n"
                       f"PDF: {paper['pdf_url']}\n"
                       f"Votes: {paper['votes']}\n"
                       f"Submitted by: {paper['submitted_by']}\n"
                       + "-" * 50
                       for paper in papers
                   ])
          )
      ]

• main.py:75-82 (registration)

  Registration of the 'get_today_papers' tool in server.list_tools(), including name, description, and input schema (no required properties).

  types.Tool(
      name="get_today_papers", 
      description="Get today's HuggingFace daily papers",
      inputSchema={
          "type": "object",
          "properties": {},
      },
  ),

• main.py:78-81 (schema)

  Input schema for 'get_today_papers' tool: empty object with no properties.

  inputSchema={
      "type": "object",
      "properties": {},
  },

• scraper.py:40-42 (helper)

  Core scraper method get_today_papers(): computes today's date and delegates to get_papers_by_date (the actual scraping logic). Called by MCP tool handler.

  def get_today_papers(self, fetch_details: bool = True) -> List[Dict]:
      today = datetime.now().strftime("%Y-%m-%d")
      return self.get_papers_by_date(today, fetch_details)

• scraper.py:19-39 (helper)

  Supporting scraper method get_papers_by_date(): fetches HTML from HF papers page, parses papers, optionally fetches details (abstract, PDF, authors from arXiv), used by get_today_papers.

  def get_papers_by_date(self, date: str, fetch_details: bool = True) -> List[Dict]:
      url = f"{self.base_url}/{date}"
      try:
          response = self.session.get(url)
          response.raise_for_status()
          papers = self._parse_papers(response.text)
          
          if fetch_details and papers:
              # 获取所有论文的详细信息，包括具体作者姓名
              for i, paper in enumerate(papers):
                  if paper.get('url'):
                      details = self._fetch_paper_details(paper['url'])
                      if details:
                          paper.update(details)
                      time.sleep(1)  # 避免请求过快
                      
          return papers
      except requests.RequestException as e:
          logging.error(f"Failed to fetch papers for {date}: {e}")
          return []