[Mediawiki-l] lucene search
pcoghlan at usa.net
Thu Aug 30 12:48:05 UTC 2007
This may not be the accepted manner of integrating Lucene but the manner in
which we are using it will, I think work on any OS that Lucene will run
We looked at the extension but as we have a Mediawiki/Drupal combo this
complicates matters in requiring a Drupal module too. It would also
complicate upgrades with the Lucene variant of both Mediawiki and Drupal
We decided instead to stream searchindex contents into an external MySQL
table. Lucene then indexes this table with no impact at all to the Mediawiki
This way we can upgrade Mediawiki or Drupal without impacting (unless
searchindex schema changes) the search. We are redirecting the search bars
and a custom advanced search page to our Lucene script to allow more control
over documents matched. We are similarly streaming searchindex content into
the same external table from Drupal meaning we have one cohesive table which
contains all data for indexing.
The table includes page_id, url, title, body and status (add, change,
delete) to help keep the index in sync with the site.
This does of course mean that we end up with search data in two places
(searchindex and Lucene Index) but I think that overhead is fairly minimal.
Once we get this launched (about a month he says optimistically) I will
submit a link. We are hopeful that the performance will be good as we now
have approaching 300k pages.
On 8/30/07 4:44 AM, "Michael Heyder" <Info at hafenvolleyballer.de> wrote:
> Hi at all,
> we wanna use the lucene search on our intranet. As I read on the
> extension docu, it's not possible to add this extension on windows systems.
> So my question: Is there a way to use this search-engine on a
> MediaWiki-l mailing list
> MediaWiki-l at lists.wikimedia.org
More information about the MediaWiki-l