Library MCP

Instructions

Implementation Reference

• main.py:428-441 (handler)

  MCP tool handler for 'get_by_tag' with @mcp.tool() decorator (registration). Includes schema via args and docstring. Executes core logic via content_manager and formats output.

  @mcp.tool()
  async def get_by_tag(tag: str, limit: int = 50) -> str:
      """Get blog content by its tag.
      
      Args:
          tag: the tag associated with content
          limit: the number of results to include
      """
      if content_manager is None:
          return "Content has not been loaded. Please ensure the server is properly initialized."
      
      matching_content = content_manager.get_by_tag(tag, limit)
      return format_content_for_output(matching_content)

• main.py:163-204 (helper)

  Core helper method in HugoContentManager that implements the tag-based search logic: normalizes tags, finds exact/partial matches case-insensitively, sorts by date descending, applies limit.

  def get_by_tag(self, tag: str, limit: int = 50) -> List[ContentFile]:
      """Find all files with a given tag"""
      matches = []
      tag_lower = tag.lower()
      
      debug_print(f"Searching for tag: '{tag_lower}'")
      for file_path, content_file in self.path_to_content.items():
          raw_tags = content_file.meta.get('tags', [])
          tags = self._normalize_tags(raw_tags)
          
          # Debug
          if tags:
              debug_print(f"File: {os.path.basename(file_path)} - Tags: {tags}")
          
          # Check for exact tag match (case insensitive)
          if any(tag_lower == t.lower() for t in tags):
              debug_print(f"Found exact tag match in {os.path.basename(file_path)}")
              matches.append(content_file)
              continue
          
          # Check if the tag is part of a tag
          for t in tags:
              if tag_lower in t.lower():
                  debug_print(f"Found partial tag match in {os.path.basename(file_path)}: '{t}'")
                  matches.append(content_file)
                  break
      
      debug_print(f"Found {len(matches)} files with tag '{tag}'")
      
      # Sort by date (most recent first)
      def get_sort_key(content_file):
          date = content_file.date
          if date is None:
              return datetime.min
          # Make date naive if it has timezone info
          if hasattr(date, 'tzinfo') and date.tzinfo is not None:
              date = date.replace(tzinfo=None)
          return date
          
      matches.sort(key=get_sort_key, reverse=True)
      
      return matches[:limit]

• main.py:372-393 (helper)

  Helper function to format the list of ContentFile objects into a readable string output for the tool response.

  def format_content_for_output(content_files: List[ContentFile]) -> str:
      """Format the content files for output"""
      if not content_files:
          return "No matching content found."
      
      result = []
      
      for i, file in enumerate(content_files):
          result.append(f"File: {file.path}")
          result.append("Metadata:")
          for key, value in file.meta.items():
              result.append(f"  {key}: {value}")
          
          # Include the full content
          result.append("Content:")
          result.append(file.data.strip())
          
          # Add separator between entries, but not after the last one
          if i < len(content_files) - 1:
              result.append("-" * 50)
      
      return "\n".join(result)