This is already implicit in the workflow, so only need to ensure it is documented. Initial deposit_new() construction is always metadata only. It is only the uploading of files that produces datapackage.json files.
There is potentially a second issue of holding a datapackage.json file, and only uploading the metadata corresponding to that, but that can and should all be directly controlled through the initial specification of client metadata anyway, and definitely doesn't need any extra parameters.
Apart from documentation, the only place this needs to be further considered will then be in the "publish" method #37, for which Figshare, for example, has an optional "is_metadata_record" parameter which should be auto-populated by deposits when "publish" is called on a deposit with no associated files.
This is already implicit in the workflow, so only need to ensure it is documented. Initial
deposit_new()construction is always metadata only. It is only the uploading of files that producesdatapackage.jsonfiles.There is potentially a second issue of holding a
datapackage.jsonfile, and only uploading the metadata corresponding to that, but that can and should all be directly controlled through the initial specification of client metadata anyway, and definitely doesn't need any extra parameters.Apart from documentation, the only place this needs to be further considered will then be in the "publish" method #37, for which Figshare, for example, has an optional "is_metadata_record" parameter which should be auto-populated by deposits when "publish" is called on a deposit with no associated files.