lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Masanz, James J." <>
Subject search influenced by token attributes
Date Thu, 15 Nov 2012 18:09:41 GMT

We have been reading that there are new flexible indexing capabilities in Lucene 4.0.  This
seems very promising and useful for what we're trying to do, but we can't find documentation
on exactly how to implement something.

Here's our problem setting: we're trying to incorporate attributes onto terms/spans in indexes,
and also in the scoring.  For example, if we're looking for a query term of "pain" but a document
says "no pain" (negation) or "severe pain" (degree) or "au bon pain" (disamiguation), we will
have indexed the document's token as having additional data.  We have implemented negation
successfully in previous versions of Lucene by using Payloads, but in principle there are
more difficult structures to match that we are not sure Payloads will suffice for.

Additionally, we want to be able to define how we score the additional attributes without
being tied to a single search model.

We've been looking at this page:
and can see some of what we're trying to do implemented through these means.  Is this the
same thing as flexible indexing, or are we missing something fundamental?

James Masanz

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message