You are viewing a plain text version of this content. The canonical link for it is here.
Posted to solr-user@lucene.apache.org by Matt Mongeau <ha...@gmail.com> on 2014/12/18 21:39:59 UTC

Multiword mispellings

Is it possible for Solr's SpellCheckComponent to suggest "Rockpoint" if the
user mistypes "Rock piont". Currently I have it making the correct
suggestions when I have "Rockpiont" or "Rock point" but not the example I
gave. Here are the relevant parts of my config files:

https://gist.github.com/halogenandtoast/c7f9335f7fa94f7b03d8

RE: Multiword mispellings

Posted by "Dyer, James" <Ja...@ingramcontent.com>.
Matt,

Unfortunately this kind of correction is not supported.  The word break spell checker works independently from the distance-based spellcheckers so it cannot correct both whitespace problems and other misspellings together.  

If you really need this, then you'll need to go with the shingle approach where you create your spellcheck field with both the base terms and also shingles (adjacent terms combined as 1 term).  In this case, "rock piont" would be considered a single term and the string difference would be 2, with one insertion (the space) and one transposition.  I believe there is a field analyzer out there that will do this for you.  I think you're supposed to set it up for both at index time (to catch when the user omits whitespace) and query time (to catch when the user adds whitespace).

James Dyer
Ingram Content Group
(615) 213-4311

-----Original Message-----
From: Matt Mongeau [mailto:halogenandtoast@gmail.com] 
Sent: Thursday, December 18, 2014 2:40 PM
To: solr-user@lucene.apache.org
Subject: Multiword mispellings

Is it possible for Solr's SpellCheckComponent to suggest "Rockpoint" if the
user mistypes "Rock piont". Currently I have it making the correct
suggestions when I have "Rockpiont" or "Rock point" but not the example I
gave. Here are the relevant parts of my config files:

https://gist.github.com/halogenandtoast/c7f9335f7fa94f7b03d8