Java CasCreationUtils.mergeTypeSystems方法代码示例

import org.apache.uima.util.CasCreationUtils; //导入方法依赖的package包/类
@Before
public void setUp() throws Exception {
    daoDesc = ExternalResourceFactory.createExternalResourceDescription(
            XmiFileTreeCorpusDAOResource.class, corpusPathString);
    tsd = CasCreationUtils
            .mergeTypeSystems(Sets.newHashSet(
                    XmiFileTreeCorpusDAO.getTypeSystem(corpusPathString),
                    TypeSystemDescriptionFactory
                            .createTypeSystemDescription(),
                    TokenizerAPI.getTypeSystemDescription(),
                    SentenceSplitterAPI.getTypeSystemDescription()));
    readerDesc = CollectionReaderFactory.createReaderDescription(
            CorpusDAOCollectionReader.class, tsd,
            CorpusDAOCollectionReader.CORPUS_DAO_KEY, daoDesc);
    CAS aCAS = CasCreationUtils.createCas(tsd, null, null, null);
    tokenizerSentenceSplitterDesc = AnalysisEngineFactory
            .createEngineDescription(Unitizer.createTokenizerSentenceSplitterAED());
    unitAnnotatorDesc = AnalysisEngineFactory.createEngineDescription(
            UnitAnnotator.class, UnitAnnotator.PARAM_UNIT_TYPE_NAMES,
            unitTypes);
}
 

import org.apache.uima.util.CasCreationUtils; //导入方法依赖的package包/类
@Before
public void setUp() throws Exception {
    daoDesc = ExternalResourceFactory.createExternalResourceDescription(
            XmiFileTreeCorpusDAOResource.class, corpusPathString);
    tsd = CasCreationUtils
            .mergeTypeSystems(Sets.newHashSet(
                    XmiFileTreeCorpusDAO.getTypeSystem(corpusPathString),
                    TypeSystemDescriptionFactory
                            .createTypeSystemDescription(),
                    TokenizerAPI.getTypeSystemDescription(),
                    SentenceSplitterAPI.getTypeSystemDescription()));
    reader = CollectionReaderFactory.createReaderDescription(
            CorpusDAOCollectionReader.class, tsd,
            CorpusDAOCollectionReader.CORPUS_DAO_KEY, daoDesc);
    CAS aCAS = CasCreationUtils.createCas(tsd, null, null, null);
    tokenizerSentenceSplitter = Unitizer.createTokenizerSentenceSplitterAED();
    unitAnnotator = AnalysisEngineFactory.createEngineDescription(
            UnitAnnotator.class, UnitAnnotator.PARAM_UNIT_TYPE_NAMES,
            unitTypes);
    unitClassifier = AnalysisEngineFactory.createEngineDescription(
            UnitClassifier.class, UnitClassifier.PARAM_CLASS_TYPE_NAMES,
            classTypes);
}
 

import org.apache.uima.util.CasCreationUtils; //导入方法依赖的package包/类
@Test
public void testTwoSentencesWithNoSpaceInBetween() throws Exception
{
    TypeSystemDescription global = TypeSystemDescriptionFactory.createTypeSystemDescription();
    TypeSystemDescription local = TypeSystemDescriptionFactory
            .createTypeSystemDescriptionFromPath(
                    "src/test/resources/desc/type/webannoTestTypes.xml");
   
    TypeSystemDescription merged = CasCreationUtils.mergeTypeSystems(asList(global, local));
    
    JCas jcas = JCasFactory.createJCas(merged);
    
    DocumentMetaData.create(jcas).setDocumentId("doc");
    jcas.setDocumentText("onetwo");
    new Token(jcas, 0, 3).addToIndexes();
    new Sentence(jcas, 0, 3).addToIndexes();
    new Token(jcas, 3, 6).addToIndexes();
    new Sentence(jcas, 3, 6).addToIndexes();
    
    writeAndAssertEquals(jcas);
}
 

import org.apache.uima.util.CasCreationUtils; //导入方法依赖的package包/类
private JCas makeJCasOneSentence() throws UIMAException
{
    TypeSystemDescription global = TypeSystemDescriptionFactory.createTypeSystemDescription();
    TypeSystemDescription local = TypeSystemDescriptionFactory
            .createTypeSystemDescriptionFromPath(
                    "src/test/resources/desc/types/webannoTestTypes.xml");
   
    TypeSystemDescription merged = CasCreationUtils.mergeTypeSystems(asList(global, local));
    
    JCas jcas = JCasFactory.createJCas(merged);
    
    DocumentMetaData.create(jcas).setDocumentId("doc");
    
    TokenBuilder<Token, Sentence> tb = new TokenBuilder<>(Token.class,
            Sentence.class);
    tb.buildTokens(jcas, "This is a test .");
    
    return jcas;
}
 

import org.apache.uima.util.CasCreationUtils; //导入方法依赖的package包/类
public static TypeSystemDescription createMultiLinkWithRoleTestTypeSytem()
    throws Exception
{
    List<TypeSystemDescription> typeSystems = new ArrayList<>();

    TypeSystemDescription tsd = new TypeSystemDescription_impl();

    // Link type
    TypeDescription linkTD = tsd.addType(LINK_TYPE, "", CAS.TYPE_NAME_TOP);
    linkTD.addFeature("role", "", CAS.TYPE_NAME_STRING);
    linkTD.addFeature("target", "", Token.class.getName());

    // Link host
    TypeDescription hostTD = tsd.addType(HOST_TYPE, "", CAS.TYPE_NAME_ANNOTATION);
    hostTD.addFeature("links", "", CAS.TYPE_NAME_FS_ARRAY, linkTD.getName(), false);

    typeSystems.add(tsd);
    typeSystems.add(TypeSystemDescriptionFactory.createTypeSystemDescription());

    return CasCreationUtils.mergeTypeSystems(typeSystems);
}
 

import org.apache.uima.util.CasCreationUtils; //导入方法依赖的package包/类
@Before
public void setUp() throws Exception {
    daoDesc = ExternalResourceFactory.createExternalResourceDescription(
            XmiFileTreeCorpusDAOResource.class, corpusPathString);
    reader = CollectionReaderFactory.createReader(
            CorpusDAOCollectionReader.class,
            XmiFileTreeCorpusDAO.getTypeSystem(corpusPathString),
            CorpusDAOCollectionReader.CORPUS_DAO_KEY, daoDesc);
    tsd = CasCreationUtils.mergeTypeSystems(Sets.newHashSet(
            XmiFileTreeCorpusDAO.getTypeSystem(corpusPathString),
            TypeSystemDescriptionFactory.createTypeSystemDescription()));
    CAS aCAS = CasCreationUtils.createCas(tsd, null, null, null);
    reader.typeSystemInit(aCAS.getTypeSystem());
}
 

import org.apache.uima.util.CasCreationUtils; //导入方法依赖的package包/类
@Before
public void setUp() throws Exception {
    daoDesc = ExternalResourceFactory.createExternalResourceDescription(
            XmiFileTreeCorpusDAOResource.class, corpusPathString);
    tsd = CasCreationUtils
            .mergeTypeSystems(Sets.newHashSet(
                    XmiFileTreeCorpusDAO.getTypeSystem(corpusPathString),
                    TypeSystemDescriptionFactory
                            .createTypeSystemDescription(),
                    TokenizerAPI.getTypeSystemDescription(),
                    SentenceSplitterAPI.getTypeSystemDescription()));
    reader = CollectionReaderFactory.createReader(
            CorpusDAOCollectionReader.class, tsd,
            CorpusDAOCollectionReader.CORPUS_DAO_KEY, daoDesc);
    CAS aCAS = CasCreationUtils.createCas(tsd, null, null, null);
    reader.typeSystemInit(aCAS.getTypeSystem());
    tokenizerSentenceSplitter = AnalysisEngineFactory
            .createEngine(Unitizer.createTokenizerSentenceSplitterAED());
    unitAnnotator = AnalysisEngineFactory.createEngine(
            UnitAnnotator.class, UnitAnnotator.PARAM_UNIT_TYPE_NAMES,
            unitTypes);
    unitClassifier = AnalysisEngineFactory.createEngine(
            UnitClassifier.class, UnitClassifier.PARAM_CLASS_TYPE_NAMES,
            classTypes);

    unitsTSV = tempFolder.newFile();
    unitsDAOWriter = AnalysisEngineFactory.createEngine(
            UnitsDAOWriter.class, UnitsDAOWriter.UNITS_TSV_PATH,
            unitsTSV.getPath());
}
 

import org.apache.uima.util.CasCreationUtils; //导入方法依赖的package包/类
private static JCas makeJCas() throws UIMAException
{
    TypeSystemDescription global = TypeSystemDescriptionFactory.createTypeSystemDescription();
    TypeSystemDescription local = TypeSystemDescriptionFactory
            .createTypeSystemDescriptionFromPath(
                    "src/test/resources/desc/type/webannoTestTypes.xml");
   
    TypeSystemDescription merged = CasCreationUtils.mergeTypeSystems(asList(global, local));
    
    JCas jcas = JCasFactory.createJCas(merged);

    DocumentMetaData.create(jcas).setDocumentId("doc");
    
    return jcas;
}
 

import org.apache.uima.util.CasCreationUtils; //导入方法依赖的package包/类
private JCas makeJCasOneSentence(String aText) throws UIMAException
{
    TypeSystemDescription global = TypeSystemDescriptionFactory.createTypeSystemDescription();
    TypeSystemDescription local = TypeSystemDescriptionFactory
            .createTypeSystemDescriptionFromPath(
                    "src/test/resources/desc/type/webannoTestTypes.xml");
   
    TypeSystemDescription merged = CasCreationUtils.mergeTypeSystems(asList(global, local));
    
    JCas jcas = JCasFactory.createJCas(merged);
    
    DocumentMetaData.create(jcas).setDocumentId("doc");
    
    TokenBuilder<Token, Sentence> tb = new TokenBuilder<>(Token.class,
            Sentence.class);
    tb.buildTokens(jcas, aText);
    
    // Remove the sentences generated by the token builder which treats the line break as a
    // sentence break
    for (Sentence s : select(jcas, Sentence.class)) {
        s.removeFromIndexes();
    }
    
    // Add a new sentence covering the whole text
    new Sentence(jcas, 0, jcas.getDocumentText().length()).addToIndexes();
    
    return jcas;
}
 

import org.apache.uima.util.CasCreationUtils; //导入方法依赖的package包/类
@Test
public void runTest() throws Exception
{
    TypeSystemDescription global = TypeSystemDescriptionFactory.createTypeSystemDescription();
    TypeSystemDescription local;
    if (new File(referenceFolder, "typesystem.xml").exists()) {
        local = TypeSystemDescriptionFactory.createTypeSystemDescriptionFromPath(
                new File(referenceFolder, "typesystem.xml").toString());
    }
    else {
        local = TypeSystemDescriptionFactory.createTypeSystemDescriptionFromPath(
                "src/test/resources/desc/type/webannoTestTypes.xml");
    }
   
    TypeSystemDescription merged = CasCreationUtils.mergeTypeSystems(asList(global, local));
    
    String targetFolder = "target/test-output/WebAnnoTsv3XReaderWriterRoundTripTest/"
            + referenceFolder.getName();
    
    CollectionReaderDescription reader = createReaderDescription(WebannoTsv3XReader.class,
            merged,
            WebannoTsv3XReader.PARAM_SOURCE_LOCATION, referenceFolder,
            WebannoTsv3XReader.PARAM_PATTERNS, "reference.tsv");
    
    AnalysisEngineDescription checker = createEngineDescription(
            DKProCoreConventionsChecker.class);
    
    AnalysisEngineDescription tsvWriter = createEngineDescription(WebannoTsv3XWriter.class,
            merged,
            WebannoTsv3XWriter.PARAM_TARGET_LOCATION, targetFolder,
            WebannoTsv3XWriter.PARAM_STRIP_EXTENSION, true);

    AnalysisEngineDescription xmiWriter = createEngineDescription(XmiWriter.class,
            merged,
            XmiWriter.PARAM_TARGET_LOCATION, targetFolder,
            XmiWriter.PARAM_STRIP_EXTENSION, true);

    SimplePipeline.runPipeline(reader, checker, tsvWriter, xmiWriter);
    
    String referenceTsv = FileUtils.readFileToString(new File(referenceFolder, "reference.tsv"),
            "UTF-8");

    String actualTsv = FileUtils.readFileToString(new File(targetFolder, "reference.tsv"),
            "UTF-8");

    //
    // The XMI files here are not compared semantically but using their serialization which
    // is subject to minor variations depending e.g. on the order in which annotation are
    // created in the CAS. Thus, this code is commented out and should only be used on a
    // case-by-case base to compare XMIs during development.
    //
    // String referenceXmi = FileUtils.readFileToString(new File(referenceFolder,
    // "reference.xmi"),
    // "UTF-8");
    //
    // String actualXmi = FileUtils.readFileToString(new File(targetFolder, "reference.xmi"),
    // "UTF-8");

    assertEquals(referenceTsv, actualTsv);
    // assertEquals(referenceXmi, actualXmi);
}
 

import org.apache.uima.util.CasCreationUtils; //导入方法依赖的package包/类
@Test
public void simpleRelGovStackedTest()
    throws Exception
{
    TypeSystemDescription customeTypesSpan = DiffUtils.createCustomTypeSystem(SPAN_TYPE,
            "webanno.custom.Multivalspan", asList("f1", "f2"), null);

    TypeSystemDescription customeTypesRel = DiffUtils.createCustomTypeSystem(RELATION_TYPE,
            "webanno.custom.Multivalrel", asList("rel1", "rel2"),
            "webanno.custom.Multivalspan");

    List<TypeSystemDescription> customTypes = new ArrayList<>();
    customTypes.add(customeTypesSpan);
    customTypes.add(customeTypesRel);
    TypeSystemDescription customType = CasCreationUtils.mergeTypeSystems(customTypes);

    Map<String, List<JCas>> casByUser = DiffUtils.loadWebAnnoTSV(customType,
            "mergecas/multivalspanrel/tale.tsv", "mergecas/multivalspanrel/tale.tsv");

    List<String> entryTypes = asList("webanno.custom.Multivalspan",
            "webanno.custom.Multivalrel");

    List<? extends DiffAdapter> diffAdapters = asList(
            new ArcDiffAdapter("webanno.custom.Multivalrel", "Dependent", "Governor", "rel1",
                    "rel2"),
            new SpanDiffAdapter("webanno.custom.Multivalspan", "f1", "f2"));

    addRandomMergeCas(casByUser);

    DiffResult result = CasDiff2.doDiff(entryTypes, diffAdapters,
            LinkCompareBehavior.LINK_TARGET_AS_LABEL, casByUser);

    JCas mergeCas = MergeCas.reMergeCas(result, getSingleCasByUser(casByUser));

    casByUser = new HashMap<>();
    JCas actual = DiffUtils.readWebAnnoTSV("mergecas/multivalspanrel/tale2.tsv", customType);
    casByUser.put("actual", asList(actual));
    casByUser.put("merge", asList(mergeCas));

    result = CasDiff2.doDiff(entryTypes, diffAdapters, LinkCompareBehavior.LINK_TARGET_AS_LABEL,
            casByUser);

    assertEquals(0, result.getDifferingConfigurationSets().size());
    assertEquals(0, result.getIncompleteConfigurationSets().size());

}
 

import org.apache.uima.util.CasCreationUtils; //导入方法依赖的package包/类
@Test
public void relStackedTest()
    throws Exception
{
    TypeSystemDescription customeTypesSpan = DiffUtils.createCustomTypeSystem(SPAN_TYPE,
            "webanno.custom.Multivalspan", asList("f1", "f2"), null);

    TypeSystemDescription customeTypesRel = DiffUtils.createCustomTypeSystem(RELATION_TYPE,
            "webanno.custom.Multivalrel", asList("rel1", "rel2"),
            "webanno.custom.Multivalspan");

    List<TypeSystemDescription> customTypes = new ArrayList<>();
    customTypes.add(customeTypesSpan);
    customTypes.add(customeTypesRel);
    TypeSystemDescription customType = CasCreationUtils.mergeTypeSystems(customTypes);

    Map<String, List<JCas>> casByUser = DiffUtils.loadXMI(customType,
            "mergecas/multivalspanrel/stackedrel1.xmi",
            "mergecas/multivalspanrel/stackedrel2.xmi");

    List<String> entryTypes = asList("webanno.custom.Multivalspan",
            "webanno.custom.Multivalrel");

    List<? extends DiffAdapter> diffAdapters = asList(
            new ArcDiffAdapter("webanno.custom.Multivalrel", "Dependent", "Governor", "rel1",
                    "rel2"),
            new SpanDiffAdapter("webanno.custom.Multivalspan", "f1", "f2"));

    addRandomMergeCas(casByUser);

    DiffResult result = CasDiff2.doDiff(entryTypes, diffAdapters,
            LinkCompareBehavior.LINK_TARGET_AS_LABEL, casByUser);

    JCas mergeCas = MergeCas.reMergeCas(result, getSingleCasByUser(casByUser));
    JCas actual = DiffUtils.readXMI("mergecas/multivalspanrel/stackedmerge.xmi", customType);

    Type relType = mergeCas.getTypeSystem().getType("webanno.custom.Multivalrel");
    int numRelMerge = CasUtil.select(mergeCas.getCas(), relType).size();
    int numRelActual = CasUtil.select(actual.getCas(), relType).size();

    Type spanType = mergeCas.getTypeSystem().getType("webanno.custom.Multivalspan");
    int numspanMerge = CasUtil.select(mergeCas.getCas(), spanType).size();
    int numspanActual = CasUtil.select(actual.getCas(), spanType).size();

    assertEquals(2, numRelMerge);
    assertEquals(2, numRelActual);

    assertEquals(4, numspanMerge);
    assertEquals(4, numspanActual);

}
 

import org.apache.uima.util.CasCreationUtils; //导入方法依赖的package包/类
@Override
public void upgradeCas(CAS aCas, SourceDocument aSourceDocument, String aUser)
    throws UIMAException, IOException
{
    TypeSystemDescription builtInTypes = TypeSystemDescriptionFactory
            .createTypeSystemDescription();
    List<TypeSystemDescription> projectTypes = getProjectTypes(aSourceDocument.getProject());
    projectTypes.add(builtInTypes);
    TypeSystemDescription allTypes = CasCreationUtils.mergeTypeSystems(projectTypes);

    // Prepare template for new CAS
    CAS newCas = JCasFactory.createJCas(allTypes).getCas();
    CASCompleteSerializer serializer = Serialization.serializeCASComplete((CASImpl) newCas);

    // Save old type system
    TypeSystem oldTypeSystem = aCas.getTypeSystem();

    // Save old CAS contents
    ByteArrayOutputStream os2 = new ByteArrayOutputStream();
    Serialization.serializeWithCompression(aCas, os2, oldTypeSystem);

    // Prepare CAS with new type system
    Serialization.deserializeCASComplete(serializer, (CASImpl) aCas);

    // Restore CAS data to new type system
    Serialization.deserializeCAS(aCas, new ByteArrayInputStream(os2.toByteArray()),
            oldTypeSystem, null);

    // Make sure JCas is properly initialized too
    aCas.getJCas();

    try (MDC.MDCCloseable closable = MDC.putCloseable(
            Logging.KEY_PROJECT_ID,
            String.valueOf(aSourceDocument.getProject().getId()))) {
        Project project = aSourceDocument.getProject();
        log.info(
                "Upgraded CAS of user [{}] for "
                        + "document [{}]({}) in project [{}]({})",
                aUser, aSourceDocument.getName(), aSourceDocument.getId(), project.getName(),
                project.getId());
    }
}
 

import org.apache.uima.util.CasCreationUtils; //导入方法依赖的package包/类
@Override
    @SuppressWarnings({ "rawtypes", "unchecked" })
    public JCas importCasFromFile(File aFile, Project aProject, String aFormat)
        throws UIMAException, IOException, ClassNotFoundException
    {
        Class readerClass = getReadableFormats().get(aFormat);
        if (readerClass == null) {
            throw new IOException("No reader available for format [" + aFormat + "]");
        }
        
        // Prepare a CAS with the project type system
        TypeSystemDescription builtInTypes = TypeSystemDescriptionFactory
                .createTypeSystemDescription();
        List<TypeSystemDescription> projectTypes = annotationService.getProjectTypes(aProject);
        projectTypes.add(builtInTypes);
        TypeSystemDescription allTypes = CasCreationUtils.mergeTypeSystems(projectTypes);
        CAS cas = JCasFactory.createJCas(allTypes).getCas();

        // Convert the source document to CAS
        CollectionReader reader = CollectionReaderFactory.createReader(readerClass,
                ResourceCollectionReaderBase.PARAM_SOURCE_LOCATION, aFile.getParentFile()
                        .getAbsolutePath(), ResourceCollectionReaderBase.PARAM_PATTERNS,
                new String[] { "[+]" + aFile.getName() });
        if (!reader.hasNext()) {
            throw new FileNotFoundException(
                    "Source file [" + aFile.getName() + "] not found in [" + aFile.getPath() + "]");
        }
        reader.getNext(cas);
        JCas jCas = cas.getJCas();

        // Create sentence / token annotations if they are missing
        boolean hasTokens = JCasUtil.exists(jCas, Token.class);
        boolean hasSentences = JCasUtil.exists(jCas, Sentence.class);

//        if (!hasTokens || !hasSentences) {
//            AnalysisEngine pipeline = createEngine(createEngineDescription(
//                    BreakIteratorSegmenter.class, 
//                    BreakIteratorSegmenter.PARAM_WRITE_TOKEN, !hasTokens,
//                    BreakIteratorSegmenter.PARAM_WRITE_SENTENCE, !hasSentences));
//            pipeline.process(jCas);
//        }
        
        if (!hasSentences) {
            splitSentences(jCas);
        }

        if (!hasTokens) {
            tokenize(jCas);
        }
        
        if (!JCasUtil.exists(jCas, Token.class) || !JCasUtil.exists(jCas, Sentence.class)) {
            throw new IOException("The document appears to be empty. Unable to detect any "
                    + "tokens or sentences. Empty documents cannot be imported.");
        }
        
        return jCas;
    }