unAPI support

nickdos
I was wondering, what is the current state of unAPI support?

What formats are supported, if any?

Does it only work in conjunction with COinS?

I've read a few posts that suggest there is _some_ unAPI support in the current version, e.g. refbase pages support it to get the abstract field into Zotero?

I've implemented a ris format unAPI on my web site but I don't get the Zotero icon in the address bar... I wanted to avoid doing COinS as my data does not fit well into COinS (patent citations).
  • noksagt
    I was wondering, what is the current state of unAPI support?
    It works.
    What formats are supported, if any?
    "mods", "marc", "endnote", "ris", "bibtex", "rdf". It takes unAPI+MODS XML from refbase (although refbase offers endnote, RIS, BibTeX, and other formats over unAPI & also has COinS).
    Does it only work in conjunction with COinS?
    It does not need COinS to work--the two are independent means of embedding metadata.
    I've read a few posts that suggest there is _some_ unAPI support in the current version, e.g. refbase pages support it to get the abstract field into Zotero?
    It gets most of the fields from refbase (not just the abstract). We'd still like to see auto-download of PDFs, but not so much as to contribute a (trivial) patch.
    I've implemented a ris format unAPI on my web site but I don't get the Zotero icon in the address bar... I wanted to avoid doing COinS as my data does not fit well into COinS (patent citations).
    You should test your unAPI implementation with other tools (there is a site to validate unAPI & there are bookmarklets. You could also look at zotero debugging. If you think your unAPI server is working & that there might be a bug in the zotero scraper, post a URL.
  • nickdos
    Thanks for those answers noksagt. I just wanted to make sure it would work before spending more time on it.

    I'm developing on a server behind a firewall, so haven't been able to run it against the unAPI validator yet. I think I'll bite the bullet and move the code up to our prod server to do that. Validator will probably find errors...

    The ris format is quite old & limited, so its good to know the other formats work too.
  • nickdos
    I've got unAPI with ris working with my site and Zotero is importing citations. Phew.

    The validator did show up errors (I actually crashed the validator the first time :-).

    Here is a URL if anyone is interested in trying it out (beta code):

    http://www.patentlens.net/patentlens_dev/patsearch.cgi?patnum=US+4904581

    The above record is the same patent that the ris documentation uses an example patent record (http://www.adeptscience.co.uk/kb/article/A626) and they match pretty well but Zotero doe not correctly import the fields.

    E.g. the patent number, application number and classification fields don't import at all and the issue date field reports both filing and publication/issue dates together. I can live with most of those problems, except for the patent number and issue date, as they are key fields.

    I presume these "issues" are a factor of Zotero being relatively young and patent citations being a less common reference type. I'll look into reporting this if its not already known.

    Which of these formats is best supported, in terms of patent field mappings:

    "mods", "marc", "endnote", "ris", "bibtex"

    In the long term I plan to use bibtools, which use mods as a common format, so maybe that's a good choice? Its just that it looks very complicated to a newbie (ris had an example patent record so was my first attempt).

    FYI, I'm prototyping in Perl but will redo in Java (Spring MVC) later.
  • noksagt
    E.g. the patent number, application number and classification fields don't import at all and the issue date field reports both filing and publication/issue dates together.
    Recently reported issue for patents in RIS.
    Which of these formats is best supported, in terms of patent field mappings
    A quick grep through the code didn't reveal any special support for any of the type, but I haven't played with it. I suspect that support will need to be added to any of the types. You can try converting your records with bibutils & prove me wrong, though. Of the formats listed, MODS XML is richer & more extensible (but I don't think there is any special support just yet).
Sign In or Register to comment.