On 8 October 2011 00:21, helix84 <heli...@centrum.sk> wrote:
> > RE #2: You could use query-time boosting to boost the relevance of the
> > 'fulltext' field... I don't have an example of the top of my head but I
> > think this should be possible... maybe even combining your original query
> > with a fulltext:{query}^10 (where '10' is the boost value... this is
> > probably too high! but suffices for demo purposes.)
>
> Thanks Kim, that's what I was looking for. Your comment about the
> boost value being too high - does it matter at all if the two OR-ed
> queries are identical?
It doesn't really matter as long as you don't plan on using boosting
for anything else... the default boost is 1.0, and you can use values
between 0 and 1 as well (eg. 0.2) so it's probably good practice to
pick something like 1.5 so that it's enough to keep the results above
the rest, but also enough wiggle room on either side should you decide
to boost some other terms/queries/fields in the future. I don't know
why I picked 10 as an example ;)
> I'll try to integrate this into the discovery search interface as
> optional checkbox, (on by default in my repository). Would others also
> be interested in this functionality? Should I submit a patch?
Sounds interesting, I'd certainly be keen to see how it optimises
searching in DSpace. A patch would be great, I'm happy to help with
testing.
Here are some Apache docs that might help:
http://wiki.apache.org/solr/SolrRelevancyCookbook
http://wiki.apache.org/solr/SolrRelevancyFAQ
http://lucene.apache.org/java/2_4_0/queryparsersyntax.html#Boosting a Term
http://wiki.apache.org/solr/SolrQuerySyntax
Note that right now, we're passing our boost in as part of the query
string Discovery uses to construct the full Solr query/url. If you
want this to be a persistent "option" when searching, and keep the
search box clean, you'll probably need to look at the way the Solr
query is constructed and stored in discovery-xmlui-api
Cheers,
Kim
------------------------------------------------------------------------------
All of the data generated in your IT infrastructure is seriously valuable.
Why? It contains a definitive record of application performance, security
threats, fraudulent activity, and more. Splunk takes this data and makes
sense of it. IT sense. And common sense.
http://p.sf.net/sfu/splunk-d2dcopy2
_______________________________________________
DSpace-tech mailing list
DSpace-tech@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/dspace-tech
