Looking at the GS primitives in the datatools.gcp module, it's not super clear that I can currently replicate all of that functionality that is provided for S3 without writing some of my own code, is that correct?