Java Code Examples for org.apache.uima.fit.factory.JCasFactory#createText()

private CAS getTestCas() throws Exception
{
    String documentText = "Dies ist ein Testtext, ach ist der schoen, der schoenste von allen"
            + " Testtexten.";
    JCas jcas = JCasFactory.createText(documentText, "de");

    NamedEntity neLabel = new NamedEntity(jcas, 0, 3);
    neLabel.setValue("LOC");
    neLabel.addToIndexes();

    // the annotation's feature value is initialized as null
    NamedEntity neNoLabel = new NamedEntity(jcas, 13, 20);
    neNoLabel.addToIndexes();

    return jcas.getCas();
}
 

@Test
public void testWriteReadExistsDeleteCas() throws Exception
{
    SourceDocument doc = makeSourceDocument(1l, 1l);
    JCas cas = JCasFactory.createText("This is a test");
    casStorageSession.add("cas", EXCLUSIVE_WRITE_ACCESS, cas.getCas());
    String user = "test";
    
    sut.writeCas(doc, cas.getCas(), user);
    assertThat(sut.getCasFile(doc, user)).exists();
    assertThat(sut.existsCas(doc, user)).isTrue();
    
    CAS cas2 = sut.readCas(doc, user);
    assertThat(cas2.getDocumentText()).isEqualTo(cas.getDocumentText());
    
    sut.deleteCas(doc, user);
    assertThat(sut.getCasFile(doc, user)).doesNotExist();
    assertThat(sut.existsCas(doc, user)).isFalse();
}
 

private List<CAS> getTestNECas(String aText, String[] aVals, int[][] aIndices) throws Exception
{
    JCas jcas = JCasFactory.createText(aText, "de");

    for (int i = 0; i < aVals.length; i++) {
        NamedEntity newNE = new NamedEntity(jcas, aIndices[i][0], aIndices[i][1]);
        newNE.setValue(aVals[i]);
        newNE.addToIndexes();
    }

    List<CAS> casses = new ArrayList<>();
    casses.add(jcas.getCas());

    return casses;
}
 

private List<CAS> getTestNECas(String aText, String[] aVals, int[][] aNEIndices,
        int[][] aSentIndices, int[][] aTokenIndices)
    throws Exception
{
    JCas jcas = JCasFactory.createText(aText, "de");

    for (int j = 0; j < aSentIndices.length; j++) {
        Sentence newSent = new Sentence(jcas, aSentIndices[j][0], aSentIndices[j][1]);
        newSent.addToIndexes();
    }

    for (int k = 0; k < aTokenIndices.length; k++) {
        Token newToken = new Token(jcas, aTokenIndices[k][0], aTokenIndices[k][1]);
        newToken.addToIndexes();
    }

    for (int i = 0; i < aVals.length; i++) {
        NamedEntity newNE = new NamedEntity(jcas, aNEIndices[i][0], aNEIndices[i][1]);
        newNE.setValue(aVals[i]);
        newNE.addToIndexes();
    }

    List<CAS> casses = new ArrayList<>();
    casses.add(jcas.getCas());

    return casses;
}
 

@Test
public void monkeyPatchTypeSystem_WithNer_CreatesScoreFeatures() throws Exception
{
    try (CasStorageSession session = CasStorageSession.open()) {
        JCas jCas = JCasFactory.createText("I am text CAS", "de");
        session.add("jCas", CasAccessMode.EXCLUSIVE_WRITE_ACCESS, jCas.getCas());
        
        when(annoService.getFullProjectTypeSystem(project))
                .thenReturn(typeSystem2TypeSystemDescription(jCas.getTypeSystem()));
        when(annoService.listAnnotationLayer(project))
                .thenReturn(asList(layer));
        doCallRealMethod().when(annoService)
                .upgradeCas(any(CAS.class), any(TypeSystemDescription.class));
        doCallRealMethod().when(annoService)
                .upgradeCas(any(CAS.class), any(CAS.class), any(TypeSystemDescription.class));

        sut.cloneAndMonkeyPatchCAS(project, jCas.getCas(), jCas.getCas());

        Type type = CasUtil.getType(jCas.getCas(), layer.getName());

        assertThat(type.getFeatures())
                .extracting(Feature::getShortName)
                .contains(feature.getName() + FEATURE_NAME_SCORE_SUFFIX)
                .contains(feature.getName() + FEATURE_NAME_SCORE_EXPLANATION_SUFFIX)
                .contains(FEATURE_NAME_IS_PREDICTION);
    }
}
 

@Test
public void testSplitSentences() throws Exception
{
    JCas jcas = JCasFactory.createText("I am one. I am two.", "en");
    
    ImportExportServiceImpl.splitSentences(jcas.getCas());
    
    assertEquals(asList("I am one.", "I am two."), toText(select(jcas, Sentence.class)));
}
 

@Test
public void testTokenize() throws Exception
{
    JCas jcas = JCasFactory.createText("i am one.i am two.", "en");
    new Sentence(jcas, 0, 9).addToIndexes();;
    new Sentence(jcas, 9, 18).addToIndexes();
    
    ImportExportServiceImpl.tokenize(jcas.getCas());
    
    assertEquals(asList("i am one.", "i am two."), toText(select(jcas, Sentence.class)));
    assertEquals(asList("i", "am", "one", ".", "i", "am", "two", "."),
            toText(select(jcas, Token.class)));
}
 

/**
 * If one annotator has provided an annotation at a given position and the other annotator did
 * not (i.e. the annotations are incomplete), then this should be detected as a disagreement. 
 */
@Test
public void thatIncompleteAnnotationIsNotMerged()
    throws Exception
{
    JCas user1 = JCasFactory.createText("word");
    token(user1, 0, 4, "X");
    
    JCas user2 = JCasFactory.createText("word");
    token(user2, 0, 4, null);
    
    Map<String, List<CAS>> casByUser = new LinkedHashMap<>();
    casByUser.put("user1", asList(user1.getCas()));
    casByUser.put("user2", asList(user2.getCas()));
    
    JCas curatorCas = createText(casByUser.values().stream()
            .flatMap(Collection::stream).findFirst().get().getDocumentText());
    
    DiffResult result = doDiff(diffAdapters, LINK_TARGET_AS_LABEL, casByUser)
            .toResult();

    sut.reMergeCas(result, document, null, curatorCas.getCas(), getSingleCasByUser(casByUser));

    assertThat(result.getDifferingConfigurationSets()).isEmpty();
    assertThat(result.getIncompleteConfigurationSets().values())
            .extracting(set -> set.getPosition())
            .usingFieldByFieldElementComparator()
            .containsExactly(new SpanPosition(null, null, 0, POS.class.getName(), 0, 4, "word",
                    null, null, -1, -1, null, null));
    
    assertThat(select(curatorCas, POS.class)).isEmpty();
}
 

/**
 * If one annotator has provided an annotation at a given position and the other annotator did
 * not (i.e. the annotations are incomplete), then this should be detected as a disagreement. 
 */
@Test
public void thatIncompleteAnnotationIsMerged()
    throws Exception
{
    JCas user1 = JCasFactory.createText("word");
    token(user1, 0, 4, "X");
    
    JCas user2 = JCasFactory.createText("word");
    token(user2, 0, 4, null);
    
    Map<String, List<CAS>> casByUser = new LinkedHashMap<>();
    casByUser.put("user1", asList(user1.getCas()));
    casByUser.put("user2", asList(user2.getCas()));
    
    JCas curatorCas = createText(casByUser.values().stream()
            .flatMap(Collection::stream).findFirst().get().getDocumentText());
    
    DiffResult result = doDiff(diffAdapters, LINK_TARGET_AS_LABEL, casByUser).toResult();

    sut.setMergeIncompleteAnnotations(true);
    sut.reMergeCas(result, document, null, curatorCas.getCas(), getSingleCasByUser(casByUser));

    assertThat(result.getDifferingConfigurationSets()).isEmpty();
    assertThat(result.getIncompleteConfigurationSets().values())
            .extracting(set -> set.getPosition())
            .usingFieldByFieldElementComparator()
            .containsExactly(new SpanPosition(null, null, 0, POS.class.getName(), 0, 4, "word",
                    null, null, -1, -1, null, null));
    
    assertThat(select(curatorCas, POS.class)).hasSize(1);
}
 

@BeforeClass
public static void setupOnce() throws Exception {
	jcas = JCasFactory.createText("test");
	fs = new Token(jcas, 0, 1);
	fs.addToIndexes();
}