You are viewing a plain text version of this content. The canonical link for it is here.
Posted to java-user@lucene.apache.org by Sam Giffney <sa...@gmail.com> on 2006/11/09 06:04:49 UTC

Google Coop - Lucene style

Most of you have probably heard of the
Google Coop Custom Search Engine
http://www.google.com/coop/cse/overview
A user defined vertical search of the Google index

The obvious and simple way to implement a Google Coop style facility
in Lucene would be to add a unique id to each document in the index
for each website selected in each users custom engine.
That's a lot of 'eaches', and it doesn't seem scalable.

Does anyone have any idea (even on a theoretical level) how Google
would be doing this and whether this kind of thing would be possible
in Lucene?

Sam

---------------------------------------------------------------------
To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-user-help@lucene.apache.org


RE: Google Coop - Lucene style

Posted by Vladimir Olenin <VO...@cihi.ca>.
I think it's pretty straighforward: the 'custom search engine' is
essentially the 'filter' that can also modify score weights of found
documents. I'd say 'coop engine' + 'your query' should be relatively
easily reducted into your 'your extended query', once you subsitute
'coop engine' with 'query parameters' you set for that engine. The only
thing that remains in the 'scorer' part of the coop engine, but that
seems relatively straightforward as well..... Or I'm missing some
crucial part altogether :).

Vlad

-----Original Message-----
From: Sam Giffney [mailto:samuelgiffney@gmail.com] 
Sent: Thursday, November 09, 2006 12:05 AM
To: java-user@lucene.apache.org
Subject: Google Coop - Lucene style

Most of you have probably heard of the
Google Coop Custom Search Engine
http://www.google.com/coop/cse/overview
A user defined vertical search of the Google index

The obvious and simple way to implement a Google Coop style facility in
Lucene would be to add a unique id to each document in the index for
each website selected in each users custom engine.
That's a lot of 'eaches', and it doesn't seem scalable.

Does anyone have any idea (even on a theoretical level) how Google would
be doing this and whether this kind of thing would be possible in
Lucene?

Sam

---------------------------------------------------------------------
To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-user-help@lucene.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-user-help@lucene.apache.org