Mnemosyne MCP

Instructions

Implementation Reference

• src/neem/mcp/tools/graph_ops.py:96-127 (handler)

  The asynchronous handler function that implements the core logic of the 'delete_graph' tool: authenticates the user, validates the graph_id parameter, submits a backend job with task_type='delete_graph', polls or streams for job completion, and returns a formatted JSON response with success status, graph_id, job_id, and result details.

  async def delete_graph_tool(
      graph_id: str,
      context: Context | None = None,
  ) -> str:
      """Delete a knowledge graph."""
      auth = MCPAuthContext.from_context(context)
      auth.require_auth()
  
      if not graph_id or not graph_id.strip():
          raise ValueError("graph_id is required and cannot be empty")
  
      metadata = await submit_job(
          base_url=backend_config.base_url,
          auth=auth,
          task_type="delete_graph",
          payload={"graph_id": graph_id.strip()},
      )
  
      if context:
          await context.report_progress(10, 100)
  
      result = await _wait_for_job_result(
          job_stream, metadata, context, auth
      )
  
      return _render_json({
          "success": True,
          "graph_id": graph_id.strip(),
          "deleted": True,
          "job_id": metadata.job_id,
          **result,
      })

• src/neem/mcp/tools/graph_ops.py:88-95 (registration)

  The @server.tool decorator that registers the 'delete_graph' tool on the FastMCP server, specifying the tool name, title, description, and input schema via function signature.

  @server.tool(
      name="delete_graph",
      title="Delete Knowledge Graph",
      description=(
          "Permanently deletes a knowledge graph and all its contents. "
          "This action cannot be undone. Use with caution."
      ),
  )

• src/neem/mcp/server/standalone_server.py:316-316 (registration)

  The call to register_graph_ops_tools which executes the tool registrations including 'delete_graph' on the MCP server instance.

  register_graph_ops_tools(mcp_server)

• src/neem/mcp/response_objects.py:570-602 (schema)

  Dataclass McpDeleteGraphResponse providing structured output schema for delete_graph operations, with fields for operation summary, deletion confirmation, backup info, and cleanup metrics.

  class McpDeleteGraphResponse(McpResponseObject):
      """Response for graph deletion operations"""
      operation_summary: McpOperationSummary
      deletion_info: Dict[str, Any]
      backup_info: Optional[Dict[str, Any]]
      cleanup_results: Dict[str, Any]
      success: bool = True
      
      @classmethod
      def from_api_response(cls, api_response: Dict[str, Any], operation_id: str, graph_id: str, user_id: Optional[str] = None) -> 'McpDeleteGraphResponse':
          """Transform API graph deletion response to MCP format"""
          import datetime
          
          return cls(
              operation_summary=McpOperationSummary(
                  operation_type="delete_graph",
                  operation_id=operation_id,
                  timestamp=datetime.datetime.now().isoformat(),
                  user_id=user_id
              ),
              deletion_info={
                  "graph_id": graph_id,
                  "deleted_at": datetime.datetime.now().isoformat(),
                  "confirmed": True
              },
              backup_info=api_response.get("backup_info"),
              cleanup_results={
                  "files_removed": api_response.get("files_removed", 0),
                  "space_freed_mb": api_response.get("space_freed_mb", 0),
                  "cache_cleared": True
              }
          )