Use Prov-o relationships to indicate the software that loaded data

[alastair]

Related to #8

In the CE Data import guidelines, section 4.6. Provenance properties, there is a section that describes how provenance of nodes should be tracked, including saying that a node was generated by something on behalf of a user.
It's unclear if we should be creating these nodes ourselves , or if the CE should do them.
Especially if if we run the importer as an algorithm, it seems like a good idea to create these nodes, but we should only do this work if we have a clear usecase for it, to avoid doing work that will be unused.