I am still wrapping my head around clowder, and I understand that dataset level extractors are new, but for my user community I'd like to be able to tell them that most of the work is to provide a script that takes a directory / filename as input and a directory for output.

With a dataset level extractor, the dataset is returned as a .zip file, this adds an extra step, but the step would presumably be the same for any dataset level extractor. So could clowder or pyclowder always make the first step when returning a dataset be to unzip (or have this as an option? Would it be possible to handle this transparently?

Also, if the dataset is on a local filesystem, is it even necessary to download + unzip, or would it be possible to simply return the path of the dataset?