[
https://issues.apache.org/jira/browse/CTAKES-67?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17652572#comment-17652572
]
Sean Finan commented on CTAKES-67:
----------------------------------
As ctakes-dictionary-lookup is no longer the default ner module (deprecated) I
changed this from a major bug to major improvement, mostly for sorting/priority
tracking.
> LuceneDictionaryImpl and iv_maxHits
> -----------------------------------
>
> Key: CTAKES-67
> URL: https://issues.apache.org/jira/browse/CTAKES-67
> Project: cTAKES
> Issue Type: Improvement
> Components: ctakes-dictionary-lookup
> Affects Versions: 3.0-incubating
> Reporter: Steven Bethard
> Priority: Major
>
> The default value for LuceneDictionaryImpl's iv_maxHits is Integer.MAX_VALUE,
> and this class uses iv_maxHits in this call:
> topDoc = iv_searcher.search(query.parse(str.replace('-', ' ')),
> iv_maxHits);
> This search method is only intended for use with a max hits up to about 100
> items:
> https://issues.apache.org/jira/browse/LUCENE-4368?focusedCommentId=13451578&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-13451578
> If you really want all the hits, then it'll speed things up (by a lot, in my
> experience) to write a simple Collector subclass that just collects them all.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
