Well, the way it works is that I put all of the LC subject authority
headings into a Solr index and my XForms application reads and parses XML
available through a URL (Solr search results). I don't have any firm ideas
how to tie relevancy in via Worldcat data, but I have a vague idea how I
will experiment with merging data from fastoutput.txt that Ralph provided
into the Solr index to make terms more relevant rather than simply ordered
alphabetically. I should be able to make the TermsComponent results come
back by each keystroke in Orbeon, but I'll have to fiddle with the js they
prepackaged in a jar file.
If you go to http://beta.scholarslab.org:9080/orbeon/ead/?guide=t1 and then
click on the tab for archival description, you can scroll down to Controlled
Access Headings. The Subject input is tied to the terms in Solr, but it's
not dynamic yet. The autosuggest options are from a static list of 250
terms from the index. Why those particular terms are shown out of hundreds
of thousands, I don't know.
Ethan
On Fri, Dec 11, 2009 at 11:36 AM, Ya'aqov Ziso <[log in to unmask]> wrote:
> Ethan, so you will be dynamcally generating XForms for each term in the
> FAST
> index? and display what the user selects? how will you make them more
> relevant?
> Could you point us to any user interface where one can see some sample
> snippets? Kindest thanks, Yašaqov
>
>
>
>
> On 12/11/09 10:46 AM, "Ethan Gruber" <[log in to unmask]> wrote:
>
> > Hi Yašaqov,
> >
> > Many of the indexes we've built over the last three years or so at the
> UVa
> > Library have been Lucene/Solr. What I am working on is actually in no
> way
> > connected to blacklight. I'm interested in using Solr with the subjects
> for
> > autosuggest because I can generate XForms instances dynamically from
> > TermsComponent.
> >
> > Ethan
>
|
