Improve COinS handling of other item types

[stakats]

COinS import should handle more fields for item types other than book and journal article. For example, blogPost has the publication title, but presently we do not grab that field for that item type.

[simon]

i don't quite understand how it would be possible to do this and stick to the COinS spec. COinS has no blogPost type, only a generic Dublin Core type, and generic Dublin Core has no publication title. unless we make some effort to extend COinS, which might not be well-liked by the people who actually created the spec, I'm not sure what there is to be done.

[stakats]

I think there is room to expand the spec – which is ad hoc, after all – but for now I was thinking of something more along the lines of exploiting those COinS parameters that seem unused. For example, maybe we could use rft.source to populate Publication Title. In other words, we could basically extend COinS by offering our own implementation/interpretation of their tags. Since right now COinS is the easiest way to add Zotero functionality to a site, the richer we can make the experience, the more quickly Z + COinS will be adopted.

[simon]

COinS is ad hoc, but OpenURL 1.0 and Dublin Core are both NISO standards. I can implement rft.source as publicationTitle (which seems justifiable, although the DCMI information is very vague), which leaves rft.format, rft.language, and rft.coverage unused.

[stakats]

We could also potentially use rft.jtitle to act as the publication title for blogPost. Right now, we just throw that field data away if we find it.

[simon]

rft.jtitle is not part of Dublin Core, and thus the OpenURL specification doesn't allow it when using the Dublin Core type. We could support it anyway, but it'd be better to rely on a spec that's better suited to carrying metadata (rather than simply identifying a resource). Hopefully, COinS will fill this place.

[simon]

(In [1118]) closes #505, Bibliography alpha sorting by case
closes #376, Bibliography export order jumbled
closes #482, Tag selector does not refresh on import/delete
closes #499, zotero RDF import of attachments has a flaw
closes #500, Improve COinS handling of other item types

• fixes an issue with importing directory hierarchy
• fixes an issue where the SpringerLink translator could fail to recognize a scrapable resource
• fixes an issue where the Nature translator could fail to retrieve an associated PDF

feel free to push the updates to the SpringerLink and Nature translators to the repository; theoretically, the RDF translator should be backwards-compatible too, but I'd like to test it with b3 before potentially breaking functionality.