Sage actually has support to directly read .d files through TimsRust. Results are slightly different than mzml as this takes the raw data directly and does different preprocesing, but your identification rates should be quite comparable whereas your analysis time is severely reduced.