[jira] [Commented] (CTAKES-383) dependency parser parses newline tokens

["ASF subversion and git services (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/CTAKES-383?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14945393#comment-14945393 ] 

ASF subversion and git services commented on CTAKES-383:
--------------------------------------------------------

Commit 1707092 from [~tmill] in branch 'ctakes/trunk'
[ https://svn.apache.org/r1707092 ]

Fixes CTAKES-383 -- filter sentence tokens for non-newline tokens.

> dependency parser parses newline tokens
> ---------------------------------------
>
>                 Key: CTAKES-383
>                 URL: https://issues.apache.org/jira/browse/CTAKES-383
>             Project: cTAKES
>          Issue Type: Bug
>          Components: ctakes-dependency-parser
>            Reporter: Tim Miller
>            Assignee: Tim Miller
>            Priority: Minor
>
> The existing sentence detector makes a hard break on newlines so the parser would never see a newline token. The new sentence detector in testing allows sentences to wrap lines and as a side effect the dependency parser and semantic role labeler are trying to parse/label the newline tokens that are now mid-sentence.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)