lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Uwe Schindler (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (SOLR-3776) Parse NamedList for JaspellLookup factory, so fuzzy and edit distance can be enabled
Date Mon, 03 Sep 2012 10:11:07 GMT

    [ https://issues.apache.org/jira/browse/SOLR-3776?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13447189#comment-13447189
] 

Uwe Schindler commented on SOLR-3776:
-------------------------------------

I did some testing on the weekend and found out that Jaspell's fuzzy algorithm is not really
useful for suggester, because it can do only prefix matching or a fuzzy match, but not combined.
This means, if you enter a prefix term, the jaspell suggest algorithm only returns results
in the ternary search tree that are exactly the distance away from the entered text.  It would
be better to match the "prefix" in a fuzzy way, but the current approach does not do this
and I see no solution how to do this.

>From my code analysis, the remaining code to do prefix matching is identical to TSTLookup,
so I wonder why we have both at all? Let's remove the Jaspell one completely, its just the
same concept with more ugly code than the straight-forward TSTLookup!

I will close this issue as "won't fix".
                
> Parse NamedList for JaspellLookup factory, so fuzzy and edit distance can be enabled
> ------------------------------------------------------------------------------------
>
>                 Key: SOLR-3776
>                 URL: https://issues.apache.org/jira/browse/SOLR-3776
>             Project: Solr
>          Issue Type: Bug
>          Components: spellchecker
>    Affects Versions: 4.0-BETA
>            Reporter: Uwe Schindler
>            Assignee: Uwe Schindler
>             Fix For: 4.0, 5.0
>
>         Attachments: SOLR-3776.patch
>
>
> Curretly JaspellLookup in Lucene already supports enabling fuzzy matches in its ternary
tree, but the factory in Solr ignores the NamedList completely and just uses "new JaspellLookup()".
This should be extended, so fuzzy suggestions would be possible (if dictionary is not too
big). This is better than no alternative until a FST-based fuzzy work -> Mike, hey! :-)

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

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


Mime
View raw message