Question

I’m looking for a way to do a simple search for a citation of a scientific paper with AJAX. Google Scholar has no API (and judging by their Terms of Use, will not have any), CiteSeerX supports the OAI-PMH API. Is anyone aware of a JavaScript / jQuery implementation that allows me to make asynchronous queries and get a list of top items? E.g. searching for knuth should yield something along the lines of

{
  {
    author: 'Knuth, DE',
    title: 'The art of computer programming. Vol. 3: sorting and searching',
    year: '1973',
    publisher: 'Addison Wesley'
  },
  {
    author: 'Graham, RL and Knuth, DE and Patashnik, O',
    title: 'Concrete mathematics: a foundation for computer science',
    year='1994',
    publisher='Addison-Wesley Reading, MA'
  },
  // ...
}

or anyone who mastered the OAI Api and knows how to wrap that into an efficient query?

Answer 1

[UPDATE]
I’ve pushed an example Mendeley search JSONP proxy to GitHub … see https://github.com/nisc/node-mendeley-search-jsonp-proxy
[/UPDATE]

I don’t have time to check out what the OAI-PMH API is and why it can’t be queried using a simple $.getJSON(), but you might also want to have a look at the Mendeley API (maybe they even implement OAI-PMH, who knows).

Particularly, the document search might be interesting.

Simply register an application (1 click). Then try something like:

# I think curl does the percent-encoding automatically. You might
# want to replace "authors:knuth" with "authors%3Aknuth", though.
curl "http://api.mendeley.com/oapi/documents/search/authors:knuth/?consumer_key=$YOUR_KEY"

I’m not sure whether you should expose your consumer key to client applications, though (better check their docs). However, you can always build a very simple API proxy (UPDATE: Considering the same origin policy, you probably have to, as they don’t seem to support JSONP or CORS).

I don’t know about their usage limits. UPDATE: 5000 searches per IP per hour.