[ 
https://issues.apache.org/jira/browse/UIMA-2234?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Peter Klügl updated UIMA-2234:
------------------------------

    Attachment: UIMA-2234.patch

Functionality added and tested with a few simple unit tests and with a bigger 
rule set.  However, it still may contain several bugs and therefore the dynamic 
anchoring is deactivated by default. The basic descriptors  of TextMarker 
projects need to be replaced and the projects needs to be cleaned.

In any case the rule engineer needs to know what he is doing when activating 
dynamic anchoring since the semantic of the rule inference changes a bit (for 
some special cases). The difference will be described in the documentation when 
there is one.

Overall, first tests with the unit test showed that dynamic anchoring can be 
200 times faster than the normal rule inference.
                
> Dynamic anchoring of rule elements
> ----------------------------------
>
>                 Key: UIMA-2234
>                 URL: https://issues.apache.org/jira/browse/UIMA-2234
>             Project: UIMA
>          Issue Type: New Feature
>          Components: TextMarker
>            Reporter: Peter Klügl
>            Assignee: Peter Klügl
>         Attachments: UIMA-2234.patch
>
>
> The rule elements currently match only in one direction: from left to right. 
> For performance reasons the start position of the rule matching should be 
> calculated by some heuristics. An example: Given a document with thousands of 
> words and with only 10 punctuation marks, and a rule "word followed by a 
> punctuation mark". That rule should not start by looking at all words of the 
> document, but at the punctuation marks. 

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: 
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira


Reply via email to