uima-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Torsten Zesch (JIRA)" <...@uima.apache.org>
Subject [jira] [Created] (UIMA-3403) selectPreceding() does not return preceding for last annotation
Date Sat, 02 Nov 2013 10:00:18 GMT
Torsten Zesch created UIMA-3403:
-----------------------------------

             Summary: selectPreceding() does not return preceding for last annotation
                 Key: UIMA-3403
                 URL: https://issues.apache.org/jira/browse/UIMA-3403
             Project: UIMA
          Issue Type: Bug
          Components: uimaFIT
            Reporter: Torsten Zesch


selectPreceding() does not return the preceding element for the last element in the CAS if
one is looking for a different annotation type than the anchor.
See the test below.
If the anchor is a token and I want the preceding tokens, everything is fine, but when I use
a token anchor to get the preceding lemmas, no lemma is returned for the last token in the
CAS.
Offset are exactly the same so the behaviour is unexpected.

@Test
    public void test() 
        throws Exception 
    {
        String testDocument = "This is a test document .\nIt is quite short .\n";

        AnalysisEngine engine = AnalysisEngineFactory.createEngine(NoOpAnnotator.class);
        JCas jcas = engine.newJCas();

        TokenBuilder<Token, Sentence> tb = TokenBuilder.create(Token.class, Sentence.class);
        tb.buildTokens(jcas, testDocument);
        engine.process(jcas);
        
        for (Token token : JCasUtil.select(jcas, Token.class)) {
            Lemma lemma = new Lemma(jcas, token.getBegin(), token.getEnd());
            lemma.addToIndexes();          
        }

        for (Token token : JCasUtil.select(jcas, Token.class)) {
//            try {
//                List<Token> previous = JCasUtil.selectPreceding(Token.class, token,
1);
//                System.out.println(previous.get(0).getCoveredText());
//            }
//            catch (IndexOutOfBoundsException e) {
//                System.out.println(token);
//            }
            try {
                List<Lemma> previous = JCasUtil.selectPreceding(Lemma.class, token,
1);
                System.out.println(previous.get(0).getCoveredText());
            }
            catch (IndexOutOfBoundsException e) {
                System.out.println(token);
            }
        }
    }




--
This message was sent by Atlassian JIRA
(v6.1#6144)

Mime
View raw message