Pydantic AI Documentation Server

Instructions

Implementation Reference

• pydantic_ai_docs_server/server.py:62-102 (handler)

  The `get_document_by_path` tool handler, implemented in `server.py` using `FastMCP`. It validates the path, resolves it to a file, and parses it using `parse_markdown_file`.

  @app.tool()
  async def get_document_by_path(path: str) -> Optional[ParsedDocument]:
      """
      Retrieves a specific document by its path relative to the Pydantic documentation root (e.g., 'usage/models.md').
      Returns the ParsedDocument if found, otherwise None.
      """
      logger.info(f"Attempting to get document by path: {path}")
      try:
          pydantic_docs_root = get_pydantic_docs_path()
          full_file_path = (pydantic_docs_root / path).resolve()
  
          if not str(full_file_path).startswith(str(pydantic_docs_root.resolve())):
              logger.warning(
                  f"Path traversal attempt or invalid path for get_document_by_path: {path}"
              )
              return None
  
          if not full_file_path.is_file():
              logger.info(
                  f"Document not found or not a file at path: {path} (resolved: {full_file_path})"
              )
              return None
  
          document = parse_markdown_file(
              file_path=full_file_path, docs_base_dir=pydantic_docs_root
          )
  
          if document:
              logger.info(f"Successfully retrieved document: {document.path}")
              return document
          else:
              logger.warning(
                  f"Failed to parse document at path: {path} (resolved: {full_file_path})"
              )
              return None
  
      except Exception as e:
          logger.error(
              f"Error in get_document_by_path for path '{path}': {e}", exc_info=True
          )
          return None