This looks great, thanks. Taking this now, though it doesn't do (almost) anything until we merge #1135

@zuphilip the container object is a workaround until the metadata schema is updated. The information comes from descriptionType seriesInformation and relatedIdentifier isPartOf. seriesInformation is a free text field, which makes it hard to consistently extract journal name, volume, issue, etc.

Ah, that's very useful though in terms of understanding what will go into the field. The way we're interpreting it will work nicely then, thanks!

DataCite JSON is also used to generate CSL-JSON. We do this in a Ruby library, but the process is simple enough to be done in another language such as Javascript. Source code at https://github.com/datacite/bolognese/blob/master/lib/bolognese/metadata_utils.rb#L115-L138

Yes, thanks -- we were aware of that (and looked at bolognese -- it's actually one of the test cases here ;) ) and that's how Zotero currently imports from Datacite, but we decided that there was some additional information in the Datacite JSON (e.g. item type for software) that would make it worthwhile to parse separately into Zotero.