Numbers in the spellchecker: Mutations based on Levenshtein distance possible?

[Christian Ortner <ch...@gmail.com>]

Hello everyone,

I'm working on a query correction feature based on collations generated by
the spellchecker. This works like a charm, except when numeric tokens are
present in the query. In that case, I don't get any corrections for the
number, although corrections for textual tokens are still made.

Any ideas?

Best regards,
Chris

Re: Numbers in the spellchecker: Mutations based on Levenshtein distance possible?

["alessandro.benedetti" <a....@sease.io>]

I would start with your spellchecker config and field analysis.
Then we can try to help you!

A shot in the dark is some kind of token filter such the word delimiter
which generates token that you are not expecting in the index ( if you are
using the Direct Spellcheck).
You can use your analysis tool in the admin dashboard or the schema browser
to take a look on what is happening.

Cheers



-----
---------------
Alessandro Benedetti
Search Consultant, R&D Software Engineer, Director
Sease Ltd. - www.sease.io
--
View this message in context: http://lucene.472066.n3.nabble.com/Numbers-in-the-spellchecker-Mutations-based-on-Levenshtein-distance-possible-tp4331635p4331646.html
Sent from the Solr - User mailing list archive at Nabble.com.