incubator-lucy-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Marvin Humphrey <>
Subject Re: [lucy-user] lucene 3.0.3
Date Tue, 02 Aug 2011 17:39:18 GMT
On Tue, Aug 02, 2011 at 09:42:08AM -0700, Eneko Agirre wrote:
> do any of you know of other Perl modules for using indexes created with
> lucene 3.0.3 or newer?

AFAIK, you have to go through a REST interface, as provided by Solr,
ElasticSearch, etc.  For local access to Lucene indexes, there is a CPAN
distro last updated in 2007, confusing named "Lucene", which actually wraps
CLucene, the C++ port of Lucene -- but CLucene does not yet offer
compatibility with Lucene 3.x.

Lucene's file format is quasi-documented, but practially speaking, it's
private.  It is so aggressively optimized, and it changes so frequently and so
dramatically that it is unrealistic for any library not derived from Java
Lucene line-by-line to keep up.  
I believe that the only library other than the direct derivatives of Lucene
which has achieved compatibility with Lucene indexes was KinoSearch, the code
base that has now become Lucy.  KinoSearch 0.05 managed compatibility with
Lucene 1.9 indexes back in 2006 -- but that was only for pure ASCII corpus
data and with no incremental indexing support. :)  We broke compatibility in
0.06 and have not looked back. 

Marvin Humphrey

View raw message