FullScope-MCP

Instructions

抓取网页内容并总结为指定比例的长度（默认20%）

Args:
    url: 要抓取和总结的网页URL
    target_ratio: 目标压缩比例，0.1-1.0之间

Returns:
    网页内容总结

Implementation Reference

• src/fullscope_mcp_server/server.py:368-399 (handler)

  The primary handler function for the 'summarize_webpage' tool. It validates parameters, scrapes the webpage using the WebScraper instance, summarizes the title and content using the ContentSummarizer instance, and returns a formatted summary with title and URL.

  @mcp.tool()
  async def summarize_webpage(url: str, ctx: Context, target_ratio: float = 0.2) -> str:
      """
      抓取网页内容并总结为指定比例的长度（默认20%）
      
      Args:
          url: 要抓取和总结的网页URL
          target_ratio: 目标压缩比例，0.1-1.0之间
      
      Returns:
          网页内容总结
      """
      try:
          # 验证参数
          if not 0.1 <= target_ratio <= 1.0:
              return "错误: target_ratio 必须在 0.1 到 1.0 之间"
          
          ctx.info(f"开始抓取并总结网页: {url}")
          
          # 先抓取网页
          title, content = await scraper.scrape_url(url)
          
          # 然后总结
          full_content = f"网页标题: {title}\n\n{content}"
          summary = await summarizer.summarize_content(full_content, target_ratio)
          
          ctx.info("网页抓取和总结完成")
          return f"网页: {url}\n标题: {title}\n\n总结:\n{summary}"
          
      except Exception as e:
          logger.error(f"网页总结失败: {e}")
          return f"网页总结失败: {str(e)}"

• src/fullscope_mcp_server/server.py:368-368 (registration)

  The @mcp.tool() decorator registers the summarize_webpage function as an MCP tool.

  @mcp.tool()

• src/fullscope_mcp_server/server.py:369-379 (schema)

  Function signature and docstring defining the input parameters (url: str, ctx: Context, target_ratio: float=0.2) and output (str), serving as the tool schema.

  async def summarize_webpage(url: str, ctx: Context, target_ratio: float = 0.2) -> str:
      """
      抓取网页内容并总结为指定比例的长度（默认20%）
      
      Args:
          url: 要抓取和总结的网页URL
          target_ratio: 目标压缩比例，0.1-1.0之间
      
      Returns:
          网页内容总结
      """

• src/fullscope_mcp_server/server.py:119-156 (helper)

  WebScraper.scrape_url: Supporting function that fetches and parses the webpage HTML using httpx and BeautifulSoup, extracts title and cleans the text content.

  async def scrape_url(self, url: str) -> tuple[str, str]:
      """
      抓取网页内容
      
      Args:
          url: 目标URL
          
      Returns:
          (title, content): 网页标题和清理后的文本内容
      """
      try:
          response = await self.session.get(url)
          response.raise_for_status()
          
          # 使用BeautifulSoup解析HTML
          soup = BeautifulSoup(response.text, 'html.parser')
          
          # 获取标题
          title = soup.find('title')
          title = title.get_text().strip() if title else "无标题"
          
          # 移除script和style标签
          for script in soup(["script", "style"]):
              script.decompose()
          
          # 提取主要内容
          content = soup.get_text()
          
          # 清理文本
          lines = (line.strip() for line in content.splitlines())
          chunks = (phrase.strip() for line in lines for phrase in line.split("  "))
          content = ' '.join(chunk for chunk in chunks if chunk)
          
          return title, content
          
      except Exception as e:
          logger.error(f"网页抓取失败 {url}: {e}")
          raise Exception(f"无法抓取网页: {str(e)}")

• src/fullscope_mcp_server/server.py:57-106 (helper)

  ContentSummarizer.summarize_content: Supporting function that uses OpenAI-compatible API (MiniMax) to summarize the provided content based on target ratio, with custom prompt option.

      async def summarize_content(self, content: str, target_ratio: float = 0.2, 
                                custom_prompt: str = None) -> str:
          """
          使用大模型总结内容
          
          Args:
              content: 要总结的内容
              target_ratio: 目标压缩比例 (默认20%)
              custom_prompt: 自定义总结提示词
          
          Returns:
              总结后的内容
          """
          try:
              # 检查内容长度，避免超出限制
              if len(content) > MAX_INPUT_TOKENS * 3:  # 粗略估算token
                  content = content[:MAX_INPUT_TOKENS * 3]
                  logger.warning("内容过长，已截断")
              
              # 构建总结提示词
              if custom_prompt:
                  prompt = custom_prompt
              else:
                  target_length = min(max(int(len(content) * target_ratio), 100), 1000)
  
                  prompt = f"""请将以下内容总结为约{target_length}字的精炼版本，保留核心信息和关键要点：
  
  {content}
  
  总结要求：
  1. 保持原文的主要观点和逻辑结构
  2. 去除冗余和次要信息
  3. 使用简洁明了的语言
  4. 确保信息的准确性和完整性"""
  
              response = self.client.chat.completions.create(
                  model=OPENAI_MODEL,
                  messages=[
                      {"role": "system", "content": "你是一个专业的内容总结专家，擅长将长文本压缩为精炼的摘要。"},
                      {"role": "user", "content": prompt}
                  ],
                  max_tokens=MAX_OUTPUT_TOKENS,
                  temperature=0.1
              )
              
              return response.choices[0].message.content.strip()
              
          except Exception as e:
              logger.error(f"内容总结失败: {e}")
              return f"总结失败: {str(e)}"