Richard Zowalla created OPENNLP-1781:
----------------------------------------
Summary: SentenceDetectorME throws StringIndexOutOfBoundsException
when sentence starts with an abbreviation
Key: OPENNLP-1781
URL: https://issues.apache.org/jira/browse/OPENNLP-1781
Project: OpenNLP
Issue Type: Bug
Affects Versions: 2.5.6
Reporter: Richard Zowalla
Assignee: Richard Zowalla
Fix For: 2.5.7, 3.0.0
When an abbreviation appears at the beginning of a sentence, OpenNLP 2.5.6's
SentenceDetectorME can throw a java.lang.StringIndexOutOfBoundsException.
This issue can be reproduced with a test like the following:
{code:java}
@Test
void testSentDetectWithAbbreviationsAtSentenceStart() {
prepareResources(true);
final String sent1 = "S. Träume sind eine Verbindung von Gedanken.";
//There is no blank space before start of the second sentence.
String[] sents = sentenceDetector.sentDetect(sent1);
double[] probs = sentenceDetector.probs();
assertAll(
() -> assertEquals(1, sents.length),
() -> assertEquals(sent1, sents[0]),
() -> assertEquals(1, probs.length));
}{code}
A practical scenario where an abbreviation might appear at the start of a
sentence is when using an ICD-10 code in a medical context.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
