lucene-solr-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Shalin Shekhar Mangar (JIRA)" <j...@apache.org>
Subject [jira] Commented: (SOLR-785) Distributed SpellCheckComponent
Date Tue, 01 Dec 2009 10:29:20 GMT

    [ https://issues.apache.org/jira/browse/SOLR-785?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12784158#action_12784158
] 

Shalin Shekhar Mangar commented on SOLR-785:
--------------------------------------------

There are a couple of problems with the current patch:
# It breaks ties using frequency info even if onlyMorePopular=false
# If more than one suggestion are returned by a shard (extendedResults=false), it uses the
first one and discards the rest
# It duplicates a lot of response writing code

Ideally, we'd like to:
# Share the same response writing code - Construct a SpellingResult using the shard responses
and pass it to the existing toNamedList method so that there is no discrepancy between distributed
and non-distributed mode
# Break ties by the configured distance measure. If distance is same and onlyMorePopular=true,
then break ties by frequency (same as how Lucene SpellChecker breaks ties)

> Distributed SpellCheckComponent
> -------------------------------
>
>                 Key: SOLR-785
>                 URL: https://issues.apache.org/jira/browse/SOLR-785
>             Project: Solr
>          Issue Type: Improvement
>          Components: spellchecker
>            Reporter: Shalin Shekhar Mangar
>            Assignee: Shalin Shekhar Mangar
>             Fix For: 1.5
>
>         Attachments: spelling-shard.patch
>
>
> Enhance the SpellCheckComponent to run in a distributed (sharded) environment.

-- 
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