storage: create new dataset to store workflow outputs using DataONE API

[mbjones]

Create a new dataset by adding new objects from the workflow, and serializing both metadata and objects and providing provenance metadata. This involves:

1. for each object, create SystemMetadata, and upload it and the object with a unique PID
2. add metadata to each object to EML metadata doc, create new version
3. add objects to data package resource map, add provenance metadata, create new version

These can be accomplished singly using atomic API operations, or using a more integrated client library like d1_python or rdataone that can do some of these steps for you.