Building a Zotero-friendly site: which format/translator?

mmulley
I'm working on a search engine for journal articles, and would love to have Zotero integration. I'd of course prefer to do the work on my side rather than having to contribute a new translator, but I'm having trouble figuring out how to provide the metadata. Could someone give me some tips? Here are my requirements:

- I need this to work on search results pages, which will generally list multiple different articles
- We provide Open Access PDFs for many articles, and so being able to provide Zotero with the URL to the full text is important
- We have relatively full metadata, including abstracts and keywords; it'd be nice to provide those
- The site is a single-page JavaScript application, so anything requiring that Zotero make HTTP requests for metadata (e.g. unapi, as I understand it) would be far more difficult to code, test, and maintain than an approach that works entirely in-browser

What are my options? I've looked at coins (where I couldn't figure out how to provide full-text links), and briefly at unapi (until I saw that it essentially required server-side rather than client-side code) and meta tags (which didn't appear to allow multiple items per page).

Thanks!
  • adamsmith
    unAPI is the only option that checks all your boxes, I'm afraid.
  • dstillman
    (Though we were going to add support for <link> auto-discovery, which would be equivalent to unAPI but much easier for publishers to implement. That hasn't happened yet, but someone could probably do it very easily.)
  • mmulley
    Ah, too bad. Is there any in-page format that can provide full-text PDF URLs, even if things like abstract/keywords aren't supported? Or is unAPI still my only option? Thanks!
  • DWL-SDCA
    edited July 18, 2018
    If you will visit my database site you can see what might be done with your site. (see: SafetyLit.org) We have freely shared our technology and scripts with other groups doing web-based bibliographic work. My email address is on the contact info page. Metadata on the site is automatically available to Zotero via the SQL/PHP & javascript- driven site (individual items or full page lists of search results). We also provide metadata in RIS, BibTeX, and EndNote format. I don't think it would be too difficult for you to adapt our code to your needs.

    As I was writing this I wonder if you are publishing original articles rather than setting up a bibliographic database. If so, the open-source Open Journal Systems is something you should explore. See: https://en.m.wikipedia.org/wiki/Open_Journal_Systems The OJS software will generate pages with metadata that is automatically available to Zotero.
  • adamsmith
    Is there any in-page format that can provide full-text PDF URLs, even if things like abstract/keywords aren't supported? Or is unAPI still my only option? Thanks!
    unAPI is currently the only option that gives you the combination of multiple import and file download.

    COinS does multiples but no files, embedded metadata does file download but not multiples.
  • DWL-SDCA
    Just to be clear, I'm offering software that can do the unAPI work you need (if your site is SQL-based) and you use your own dedicated server.

This is an old discussion that has not been active in a long time. Before commenting here, you should strongly consider starting a new discussion instead. If you think the content of this discussion is still relevant, you can link to it from your new discussion.

Sign In or Register to comment.