lucene-solr-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Michael Ludwig (JIRA)" <j...@apache.org>
Subject [jira] Commented: (SOLR-1204) Enhance SpellingQueryConverter to handle UTF-8 instead of ASCII only
Date Fri, 05 Jun 2009 12:22:07 GMT

    [ https://issues.apache.org/jira/browse/SOLR-1204?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12716594#action_12716594
] 

Michael Ludwig commented on SOLR-1204:
--------------------------------------

Sorry for being inexact in my last comment. I wrote "any valid UTF-8 character" when I meant
"any valid UTF-8 letter", which is what the patch contains, and which isn't equivalent to
what I wrote.

In order to produce a correct patch, I need to know what are legal field names. It can hardly
be "any UTF-8 string" as that will also contain the colon, which is already used to delimit
field names from query strings. What about digits? Asterisk? Dash (minus)? Underscore? Space?
Tabulator?

My original idea was to allow letters, digits and underscore. But this may not be sufficient.

> Enhance SpellingQueryConverter to handle UTF-8 instead of ASCII only
> --------------------------------------------------------------------
>
>                 Key: SOLR-1204
>                 URL: https://issues.apache.org/jira/browse/SOLR-1204
>             Project: Solr
>          Issue Type: Improvement
>          Components: spellchecker
>    Affects Versions: 1.3
>            Reporter: Michael Ludwig
>            Assignee: Shalin Shekhar Mangar
>            Priority: Trivial
>             Fix For: 1.4
>
>         Attachments: SpellingQueryConverter.java.diff, SpellingQueryConverter.java.diff
>
>
> Solr - User - SpellCheckComponent: queryAnalyzerFieldType
> http://www.nabble.com/SpellCheckComponent%3A-queryAnalyzerFieldType-td23870668.html
> In the above thread, it was suggested to extend the SpellingQueryConverter to cover the
full UTF-8 range instead of handling US-ASCII only. This might be as simple as changing the
regular expression used to tokenize the input string to accept a sequence of one or more Unicode
letters ( \p{L}+ ) instead of a sequence of one or more word characters ( \w+ ).
> See http://java.sun.com/j2se/1.4.2/docs/api/java/util/regex/Pattern.html for Java regular
expression reference.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message