[
https://issues.apache.org/jira/browse/OPENNLP-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16447772#comment-16447772
]
ASF GitHub Bot commented on OPENNLP-1194:
-----------------------------------------
kottmann commented on a change in pull request #312: OPENNLP-1194: Adds type
name filter to BratDocumentParser
URL: https://github.com/apache/opennlp/pull/312#discussion_r183316113
##########
File path:
opennlp-tools/src/main/java/opennlp/tools/formats/brat/BratDocumentParser.java
##########
@@ -145,5 +148,13 @@ public BratDocumentParser(SentenceDetector
sentenceDetector, Tokenizer tokenizer
return samples;
}
+
+ private boolean isSpanAnnotation(BratAnnotation ann) {
+ if (ann instanceof SpanAnnotation &&
Review comment:
Probably better to be clear about what we expect:
- Don't enable name type filtering, in that case pass null as nameTypes
parameter
- If an empty set is passed an IllegalArgumentException could be thrown,
doesn't make sense to filter all
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
> Brat Document Parser should support name type filters
> -----------------------------------------------------
>
> Key: OPENNLP-1194
> URL: https://issues.apache.org/jira/browse/OPENNLP-1194
> Project: OpenNLP
> Issue Type: Improvement
> Components: Formats
> Reporter: William Colen
> Assignee: William Colen
> Priority: Major
> Fix For: 1.8.5
>
>
> Brat Document Parser fails if there is a span overlap. Sometimes we are
> interested in only some types. In that case we could ignore the overlapping
> that are not of our interest.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
