com.optimaize.langdetect.profiles.LanguageProfileReader Java Examples

static private void initOptimaize() {
    synchronized (initGuard) {
        if ((textObjectFactory != null) && (languageDetector != null)) return;

        // origin: https://github.com/optimaize/language-detector
        // load all languages:
        List<LanguageProfile> languageProfiles;
        try {
            languageProfiles = new LanguageProfileReader().readAllBuiltIn();
        } catch (IOException e) {
            throw new UncheckedIOException(e);
        }

        //build language detector:
        languageDetector = LanguageDetectorBuilder.create(NgramExtractors.standard())
                                                  .withProfiles(languageProfiles)
                                                  .build();

        //create a text object factory
        textObjectFactory = CommonTextObjectFactories.forDetectingOnLargeText();
    }
}
 

private static LanguageDetector getLanguageDetector(){
    if (languageDetector == null){
        try {
            List<String> languages = new ArrayList<>();
            for(Language lg : Language.values())
                languages.add(lg.getAbrev().toLowerCase());

            List<LanguageProfile> languageProfiles = new LanguageProfileReader().read(languages);
            languageDetector = LanguageDetectorBuilder.create(NgramExtractors.standard())
                            .withProfiles(languageProfiles).build();
        }
        catch (IOException e) {
            LOG.error("Translator.getLanguageDetector", e);
        }
    }
    return languageDetector;
}
 

protected final LanguageDetector getLanguageDetector() {
    if (detectorInstance == null) {
        synchronized (AbstractOptimaizeFilter.class) {
            if (detectorInstance == null) {
                try {
                    detectorInstance = LanguageDetectorBuilder.create(NgramExtractors.standard())
                            .shortTextAlgorithm(0)
                            .withProfiles(new LanguageProfileReader().readAllBuiltIn())
                            .build();
                } catch (IOException e) {
                    throw new RuntimeIOException(e);
                }
            }
        }
    }

    return detectorInstance;
}
 

/**
 * Using all language profiles from the given directory.
 */
private LanguageDetector makeDetector() throws IOException {
    double alpha = getParamDouble("alpha", DEFAULT_ALPHA);
    String profileDirectory = requireParamString("directory") + "/";
    Optional<Long> seed = Optional.fromNullable(getParamLongOrNull("seed"));

    List<LanguageProfile> languageProfiles = new LanguageProfileReader().readAll(new File(profileDirectory));

    return LanguageDetectorBuilder.create(NgramExtractors.standard()).alpha(alpha).seed(seed).shortTextAlgorithm(50).withProfiles(languageProfiles).build();
}
 

public LanguageCheckerImpl() {
    //build language detector:
    try {
        languageDetector = LanguageDetectorBuilder.create(NgramExtractors.standard())
                .withProfiles(new LanguageProfileReader().readAllBuiltIn()) //load all languages:
                .build();
    } catch (IOException e) {
        log.error("LanguageCheckerImpl {}", e.getMessage());
    }

    //create a text object factory
    textObjectFactory = CommonTextObjectFactories.forDetectingOnLargeText();
}
 

public NonEnglishCheckerService() {
    try {
        languageProfiles = new LanguageProfileReader().readAllBuiltIn();
        optimaizeDetector = LanguageDetectorBuilder.create(NgramExtractors.standard())
                .withProfiles(languageProfiles)
                .build();
        textObjectFactory = CommonTextObjectFactories.forDetectingOnLargeText();
        tikaDetector = new OptimaizeLangDetector().loadModels();
        writer = new LanguageWriter(tikaDetector);

    } catch (IOException e) {
        e.printStackTrace();
    }
}
 

private LanguageDetectionSingleton() throws IOException {
	// load all languages
	List<LanguageProfile> languageProfiles = new LanguageProfileReader().readAllBuiltIn();

	// build language detector
	languageDetector = LanguageDetectorBuilder.create(NgramExtractors.standard())
			.withProfiles(languageProfiles)
			.build();

	// create a text object factory
	textObjectFactory = CommonTextObjectFactories.forDetectingShortCleanText();
}
 

@Override
public void doInitialize(UimaContext aContext) throws ResourceInitializationException {
  try {
    List<LanguageProfile> languageProfiles = new LanguageProfileReader().readAllBuiltIn();
    languageDetector =
        LanguageDetectorBuilder.create(NgramExtractors.standard())
            .withProfiles(languageProfiles)
            .build();

    textObjectFactory = CommonTextObjectFactories.forDetectingOnLargeText();
  } catch (IOException ioe) {
    throw new ResourceInitializationException(ioe);
  }
}
 

/**
 * Using all language profiles from the given directory.
 */
private LanguageDetector makeDetector() throws IOException {
    double alpha = getParamDouble("alpha", DEFAULT_ALPHA);
    String profileDirectory = requireParamString("directory") + "/";
    Optional<Long> seed = Optional.fromNullable(getParamLongOrNull("seed"));

    List<LanguageProfile> languageProfiles = new LanguageProfileReader().readAll(new File(profileDirectory));

    return LanguageDetectorBuilder.create(NgramExtractors.standard())
            .alpha(alpha)
            .seed(seed)
            .shortTextAlgorithm(50)
            .withProfiles(languageProfiles)
            .build();
}
 

public DataLanguageDetectorImplTest() throws IOException {
    List<LanguageProfile> languageProfiles = new LanguageProfileReader().readAllBuiltIn();

    shortDetector = LanguageDetectorBuilder.create(NgramExtractors.standard())
            .shortTextAlgorithm(100)
            .withProfiles(languageProfiles)
            .build();

    longDetector = LanguageDetectorBuilder.create(NgramExtractors.standard())
            .shortTextAlgorithm(0)
            .withProfiles(new LanguageProfileReader().readAllBuiltIn())
            .build();
}
 

/**
 * Adds all {@link LanguageProfile}s that are available when calling {@link LanguageProfileReader#readAllBuiltIn()}.
 */
public LanguageProfileValidator loadAllBuiltInLanguageProfiles() throws IOException {
    this.languageProfiles.addAll(new LanguageProfileReader().readAllBuiltIn());
    return this;
}
 

private static NgramFrequencyData forAll(int gramSize) throws IOException {
    List<LanguageProfile> languageProfiles = new LanguageProfileReader().readAllBuiltIn();
    return NgramFrequencyData.create(languageProfiles, ImmutableSet.of(gramSize));
}
 

public OptimaizeLanguageGuesser() throws IOException {
    this.languageDetector = LanguageDetectorBuilder.create(NgramExtractors.standard())
                    .withProfiles(new LanguageProfileReader().readAllBuiltIn())
                    .build();
}
 

/**
 * Adds all {@link LanguageProfile}s that are available when calling {@link LanguageProfileReader#readAllBuiltIn()}.
 */
public LanguageProfileValidator loadAllBuiltInLanguageProfiles() throws IOException {
    this.languageProfiles.addAll(new LanguageProfileReader().readAllBuiltIn());
    return this;
}
 

private static NgramFrequencyData forAll(int gramSize) throws IOException {
    List<LanguageProfile> languageProfiles = new LanguageProfileReader().readAllBuiltIn();
    return NgramFrequencyData.create(languageProfiles, ImmutableSet.of(gramSize));
}
 

public DataLanguageDetectorImplTest() throws IOException {
    List<LanguageProfile> languageProfiles = new LanguageProfileReader().readAllBuiltIn();

    shortDetector = LanguageDetectorBuilder.create(NgramExtractors.standard()).shortTextAlgorithm(100).withProfiles(languageProfiles).build();

    longDetector = LanguageDetectorBuilder.create(NgramExtractors.standard()).shortTextAlgorithm(0).withProfiles(new LanguageProfileReader().readAllBuiltIn()).build();
}