Thursday, 30 August 2007

Piggy Bank and bioGUID - browsing the biodiversity Semantic Web

Piggy Bank is, according to its developers:
...a Firefox extension that turns your browser into a mashup platform, by allowing you to extract data from different web sites and mix them together.

I've been rather underwhelmed in the past, but Jim Balhoff asked me why bioGUID didn't play ball, so I decided to take another look.

On pointing Piggy Bank at a bioGUID page, such as casent0498428 and clicking on the "data coin" icon in the bottom right corner of the Firefox browser window failed to produce anything interesting (Piggy Bank just grabbed some text from the web page, not the underlying RDF). After a bit of fussing, I finally got bioGUID to work with Piggy Bank.

BioGUID resolves a GUID and returns RDF as an XML document, with a XSL instruction included so that the web browser renders it as HTML. This means that a user with a web browser gets a nicley formatted page, but a Semantic Web tool gets RDF. However, Piggy Bank needs some help. First, I include a link to the RDF in the <head> element of the HTML, as described on the Piggy Bank web sitee.g.:

<link rel="alternate" type="application/rdf+xml"

The &format=raw parameter is important as this tells bioGUID to return RDF with the MIME type "application/rdf+xml". Piggy Bank needs this MIME type set, otherwise it just parses HTML. I normally return RDF as "application/xml" so it displays nicely in web browsers if I'm debugging, and to ensure the XML is transformed into HTML. Clearly I'm going to have to rethink this, perhaps by exploring content negotiation.

Anyway, for the time being Piggy Bank and bioGUID seem to work together, as this screen shot shows:

If you want to explore Piggy Bank as a Semantic WEb browser for biodoversity data, one place to start is and the sequences it links to, many of which link to specimens.

Wednesday, 15 August 2007

OpenURL and COinS

I've improved the bioGUID OpenURL service, based on my experience with the OpenURL Referrer extension for FireFox. This extension takes COinS links in HTML and inserts a link to an OpenURL resolver. The EDIT scratchpads being developed by Vince Smith and colleagues at the Natural History Museum contain COinS (see the APOL - Abyssal Polychaetes OnLine bibliography, for an example). I installed the OpenURL Referrer extension, but my OpenURL resolver broke. Turned out the the COinS want OpenURL 1.0, whereas I supported the much simpler version 0.1 (Caveat Lector pretty much sums up how I feel about this). Anyway, a quick hack of my code and it now works.

If you add as a profile, you can then use my OpenURL resolver (there's even a little button to click on at Once installed, scratchpads such as APOL - Abyssal Polychaetes OnLine bibliography now have links to the OpenURL resolver.

Why use bioGUID's service? Well, in addition to supporting papers that have DOIs, I'm adding other sources, as mentioned earlier in this blog, and on iPhylo. Some 9000 freely available full text articles are now available through this service. There's still a lot more to do to this service, such as supporting other identifiers (e.g., PubMed and SICIs).

Thanks to Julius Welby for reminding me about COinS.