You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@uima.apache.org by "John (Jira)" <de...@uima.apache.org> on 2021/09/24 14:18:00 UTC

[jira] [Commented] (UIMA-2761) Ruta: Rule matching on annotations with different boundary visibility

    [ https://issues.apache.org/jira/browse/UIMA-2761?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17419806#comment-17419806 ] 

John commented on UIMA-2761:
----------------------------

Similar problem if match starts with invisible.

Text: “UIMA RUTA RULEZ”
 {code}
ADDRETAINTYPE(SPACE);
SPACE "RUTA" SPACE { -> MARK(Annotation1, 2, 2), MARK(Annotation2, 1, 2), MARK(Annotation3, 2, 3)};
REMOVERETAINTYPE(SPACE);
Annotation1; // Fires
Annotation2; // Doesn't fire
Annotation3; // Doesn't fire
{code}

So all annotation must be trimmed to match as expected.

> Ruta: Rule matching on annotations with different boundary visibility
> ---------------------------------------------------------------------
>
>                 Key: UIMA-2761
>                 URL: https://issues.apache.org/jira/browse/UIMA-2761
>             Project: UIMA
>          Issue Type: Bug
>          Components: Ruta
>            Reporter: Peter Klügl
>            Assignee: Peter Klügl
>            Priority: Major
>
> There are actually two problems with annotations with different boundary visibility. Annotations like Name (covered text "Peter ") where (with default settings) the begin anchor is visible, but the end anchor is invisible (because of the space).
> The first problem is that the direction in which the rule elements match using dynamic anchoring is important, if any results are returned. An example: The rule "CW Name;" matches differently depending on the rule element which starts the match.
> The second problem is that there is additionally a bug in finding the next matching position, if the current match ends with something invisible.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)