Examples with TextAnnotation com.google.cloud.vision.v1.TextAnnotation used on opensource projects

Example 21 with TextAnnotation

use of com.google.cloud.vision.v1.TextAnnotation in project kie-wb-common by kiegroup.

the class DMNMarshaller method marshall.

@Override
public String marshall(final Diagram<Graph, Metadata> diagram) throws IOException {
    Graph<?, Node<View, ?>> g = diagram.getGraph();
    Map<String, org.kie.dmn.model.v1_1.DRGElement> nodes = new HashMap<>();
    Map<String, org.kie.dmn.model.v1_1.TextAnnotation> textAnnotations = new HashMap<>();
    @SuppressWarnings("unchecked") Node<View<DMNDiagram>, ?> dmnDiagramRoot = (Node<View<DMNDiagram>, ?>) findDMNDiagramRoot(g);
    Definitions definitionsStunnerPojo = dmnDiagramRoot.getContent().getDefinition().getDefinitions();
    org.kie.dmn.model.v1_1.Definitions definitions = DefinitionsConverter.dmnFromWB(definitionsStunnerPojo);
    if (definitions.getExtensionElements() == null) {
        definitions.setExtensionElements(new org.kie.dmn.model.v1_1.Definitions.ExtensionElements());
    }
    org.kie.workbench.common.dmn.backend.definition.v1_1.dd.DMNDiagram dmnDDDMNDiagram = new org.kie.workbench.common.dmn.backend.definition.v1_1.dd.DMNDiagram();
    definitions.getExtensionElements().getAny().add(dmnDDDMNDiagram);
    for (Node<?, ?> node : g.nodes()) {
        if (node.getContent() instanceof View<?>) {
            View<?> view = (View<?>) node.getContent();
            if (view.getDefinition() instanceof DRGElement) {
                DRGElement n = (org.kie.workbench.common.dmn.api.definition.v1_1.DRGElement) view.getDefinition();
                nodes.put(n.getId().getValue(), stunnerToDMN(node));
                dmnDDDMNDiagram.getAny().add(stunnerToDDExt((View<? extends DMNElement>) view));
            } else if (view.getDefinition() instanceof TextAnnotation) {
                TextAnnotation textAnnotation = (TextAnnotation) view.getDefinition();
                textAnnotations.put(textAnnotation.getId().getValue(), textAnnotationConverter.dmnFromNode((Node<View<TextAnnotation>, ?>) node));
                dmnDDDMNDiagram.getAny().add(stunnerToDDExt((View<? extends DMNElement>) view));
                List<org.kie.dmn.model.v1_1.Association> associations = AssociationConverter.dmnFromWB((Node<View<TextAnnotation>, ?>) node);
                definitions.getArtifact().addAll(associations);
            }
        }
    }
    nodes.values().forEach(definitions.getDrgElement()::add);
    textAnnotations.values().forEach(definitions.getArtifact()::add);
    String marshalled = marshaller.marshal(definitions);
    return marshalled;
}

Example 22 with TextAnnotation

use of com.google.cloud.vision.v1.TextAnnotation in project opencast by opencast.

the class Mpeg7CaptionConverter method exportCaption.

@Override
public void exportCaption(OutputStream outputStream, List<Caption> captions, String language) throws IOException {
    Mpeg7Catalog mpeg7 = Mpeg7CatalogImpl.newInstance();
    MediaTime mediaTime = new MediaTimeImpl(0, 0);
    Audio audioContent = mpeg7.addAudioContent("captions", mediaTime, null);
    @SuppressWarnings("unchecked") TemporalDecomposition<AudioSegment> captionDecomposition = (TemporalDecomposition<AudioSegment>) audioContent.getTemporalDecomposition();
    int segmentCount = 0;
    for (Caption caption : captions) {
        // Get all the words/parts for the transcript
        String[] words = caption.getCaption();
        if (words.length == 0)
            continue;
        // Create a new segment
        AudioSegment segment = captionDecomposition.createSegment("segment-" + segmentCount++);
        Time captionST = caption.getStartTime();
        Time captionET = caption.getStopTime();
        // Calculate start time
        Calendar startTime = Calendar.getInstance(TimeZone.getTimeZone("UTC"));
        startTime.setTimeInMillis(0);
        startTime.add(Calendar.HOUR_OF_DAY, captionST.getHours());
        startTime.add(Calendar.MINUTE, captionST.getMinutes());
        startTime.add(Calendar.SECOND, captionST.getSeconds());
        startTime.add(Calendar.MILLISECOND, captionST.getMilliseconds());
        // Calculate end time
        Calendar endTime = Calendar.getInstance(TimeZone.getTimeZone("UTC"));
        endTime.setTimeInMillis(0);
        endTime.add(Calendar.HOUR_OF_DAY, captionET.getHours());
        endTime.add(Calendar.MINUTE, captionET.getMinutes());
        endTime.add(Calendar.SECOND, captionET.getSeconds());
        endTime.add(Calendar.MILLISECOND, captionET.getMilliseconds());
        long startTimeInMillis = startTime.getTimeInMillis();
        long endTimeInMillis = endTime.getTimeInMillis();
        long duration = endTimeInMillis - startTimeInMillis;
        segment.setMediaTime(new MediaTimeImpl(startTimeInMillis, duration));
        TextAnnotation textAnnotation = segment.createTextAnnotation(0, 0, language);
        // Collect all the words in the segment
        StringBuffer captionLine = new StringBuffer();
        // Add each words/parts as segment to the catalog
        for (String word : words) {
            if (captionLine.length() > 0)
                captionLine.append(' ');
            captionLine.append(word);
        }
        // Append the text to the annotation
        textAnnotation.addFreeTextAnnotation(new FreeTextAnnotationImpl(captionLine.toString()));
    }
    Transformer tf = null;
    try {
        tf = TransformerFactory.newInstance().newTransformer();
        DOMSource xmlSource = new DOMSource(mpeg7.toXml());
        tf.transform(xmlSource, new StreamResult(outputStream));
    } catch (TransformerConfigurationException e) {
        logger.warn("Error serializing mpeg7 captions catalog: {}", e.getMessage());
        throw new IOException(e);
    } catch (TransformerFactoryConfigurationError e) {
        logger.warn("Error serializing mpeg7 captions catalog: {}", e.getMessage());
        throw new IOException(e);
    } catch (TransformerException e) {
        logger.warn("Error serializing mpeg7 captions catalog: {}", e.getMessage());
        throw new IOException(e);
    } catch (ParserConfigurationException e) {
        logger.warn("Error serializing mpeg7 captions catalog: {}", e.getMessage());
        throw new IOException(e);
    }
}

Example 23 with TextAnnotation

use of com.google.cloud.vision.v1.TextAnnotation in project opencast by opencast.

the class SolrIndexManager method importantKeywordsString.

/**
 * Generates a string with the most important kewords from the text annotation.
 *
 * @param sortedAnnotations
 * @return The keyword string.
 */
static StringBuffer importantKeywordsString(SortedSet<TextAnnotation> sortedAnnotations) {
    // important keyword:
    // - high relevance
    // - high confidence
    // - occur often
    // - more than MAX_CHAR chars
    // calculate keyword occurences (histogram) and importance
    ArrayList<String> list = new ArrayList<String>();
    Iterator<TextAnnotation> textAnnotations = sortedAnnotations.iterator();
    TextAnnotation textAnnotation = null;
    String keyword = null;
    HashMap<String, Integer> histogram = new HashMap<String, Integer>();
    HashMap<String, Double> importance = new HashMap<String, Double>();
    int occ = 0;
    double imp;
    while (textAnnotations.hasNext()) {
        textAnnotation = textAnnotations.next();
        Iterator<KeywordAnnotation> keywordAnnotations = textAnnotation.keywordAnnotations();
        while (keywordAnnotations.hasNext()) {
            KeywordAnnotation annotation = keywordAnnotations.next();
            keyword = annotation.getKeyword().toLowerCase();
            if (keyword.length() > MAX_CHAR) {
                occ = 0;
                if (histogram.keySet().contains(keyword)) {
                    occ = histogram.get(keyword);
                }
                histogram.put(keyword, occ + 1);
                // here the importance value is calculated
                // from relevance, confidence and frequency of occurence.
                imp = (RELEVANCE_BOOST * getMaxRelevance(keyword, sortedAnnotations) + getMaxConfidence(keyword, sortedAnnotations)) * (occ + 1);
                importance.put(keyword, imp);
            }
        }
    }
    // get the MAX_IMPORTANT_COUNT most important keywords
    StringBuffer buf = new StringBuffer();
    while (list.size() < MAX_IMPORTANT_COUNT && importance.size() > 0) {
        double max = 0.0;
        String maxKeyword = null;
        // get maximum from importance list
        for (Entry<String, Double> entry : importance.entrySet()) {
            keyword = entry.getKey();
            if (max < entry.getValue()) {
                max = entry.getValue();
                maxKeyword = keyword;
            }
        }
        // pop maximum
        importance.remove(maxKeyword);
        // append keyword to string
        if (buf.length() > 0)
            buf.append(" ");
        buf.append(maxKeyword);
    }
    return buf;
}

Example 24 with TextAnnotation

use of com.google.cloud.vision.v1.TextAnnotation in project opencast by opencast.

the class SolrIndexManager method getMaxConfidence.

/**
 * Gets the maximum confidence for a given keyword in the text annotation.
 *
 * @param keyword
 * @param sortedAnnotations
 * @return The maximum confidence value.
 */
static double getMaxConfidence(String keyword, SortedSet<TextAnnotation> sortedAnnotations) {
    double max = 0.0;
    String needle = null;
    TextAnnotation textAnnotation = null;
    Iterator<TextAnnotation> textAnnotations = sortedAnnotations.iterator();
    while (textAnnotations.hasNext()) {
        textAnnotation = textAnnotations.next();
        Iterator<KeywordAnnotation> keywordAnnotations = textAnnotation.keywordAnnotations();
        while (keywordAnnotations.hasNext()) {
            KeywordAnnotation ann = keywordAnnotations.next();
            needle = ann.getKeyword().toLowerCase();
            if (keyword.equals(needle)) {
                if (max < textAnnotation.getConfidence()) {
                    max = textAnnotation.getConfidence();
                }
            }
        }
    }
    return max;
}