Examples with GroupOfShapes com.joliciel.jochre.graphics.GroupOfShapes used on opensource projects

Example 16 with GroupOfShapes

use of com.joliciel.jochre.graphics.GroupOfShapes in project jochre by urieli.

the class MergeEvaluator method evaluate.

public FScoreCalculator<String> evaluate(JochreCorpusGroupReader groupReader, ShapeMerger shapeMerger) {
    LOG.debug("evaluate");
    FScoreCalculator<String> fScoreCalculator = new FScoreCalculator<String>();
    while (groupReader.hasNext()) {
        GroupOfShapes group = groupReader.next();
        Shape previousShape = null;
        for (Shape shape : group.getShapes()) {
            if (previousShape != null) {
                ShapePair mergeCandidate = new ShapePair(previousShape, shape);
                double widthRatio = 0;
                double distanceRatio = 0;
                if (mergeCandidate.getXHeight() > 0) {
                    widthRatio = (double) mergeCandidate.getWidth() / (double) mergeCandidate.getXHeight();
                    distanceRatio = (double) mergeCandidate.getInnerDistance() / (double) mergeCandidate.getXHeight();
                }
                boolean shouldMerge = false;
                if (mergeCandidate.getFirstShape().getLetter().startsWith("|")) {
                    if (mergeCandidate.getSecondShape().getLetter().length() == 0 || mergeCandidate.getSecondShape().getLetter().endsWith("|"))
                        shouldMerge = true;
                } else if (mergeCandidate.getSecondShape().getLetter().endsWith("|")) {
                    if (mergeCandidate.getFirstShape().getLetter().length() == 0)
                        shouldMerge = true;
                }
                if (LOG.isTraceEnabled()) {
                    LOG.trace(mergeCandidate.toString());
                    LOG.trace("widthRatio: " + widthRatio);
                    LOG.trace("distanceRatio: " + distanceRatio);
                    LOG.trace("shouldMerge: " + shouldMerge);
                }
                if (widthRatio <= maxWidthRatio && distanceRatio <= maxDistanceRatio) {
                    double mergeProb = shapeMerger.checkMerge(previousShape, shape);
                    boolean wantsToMerge = (mergeProb >= minProbabilityForDecision);
                    fScoreCalculator.increment(shouldMerge ? "YES" : "NO", wantsToMerge ? "YES" : "NO");
                } else {
                    LOG.trace("too wide");
                    if (shouldMerge)
                        fScoreCalculator.increment("YES", "WIDE");
                    else
                        fScoreCalculator.increment("NO", "NO");
                }
            // too wide?
            }
            // have previous shape?
            previousShape = shape;
        }
    // next shape
    }
    // next group
    return fScoreCalculator;
}

Example 17 with GroupOfShapes

use of com.joliciel.jochre.graphics.GroupOfShapes in project jochre by urieli.

the class BeamSearchImageAnalyser method analyseInternal.

public void analyseInternal(JochreImage image) {
    LOG.debug("Analysing image " + image.getId());
    if (currentMonitor != null) {
        currentMonitor.setCurrentAction("imageMonitor.analysingImage", new Object[] { image.getPage().getIndex() });
    }
    for (LetterGuessObserver observer : observers) {
        observer.onImageStart(image);
    }
    if (totalShapeCount < 0)
        totalShapeCount = image.getShapeCount();
    for (Paragraph paragraph : image.getParagraphs()) {
        LOG.debug("Analysing paragraph " + paragraph.getIndex() + " (id=" + paragraph.getId() + ")");
        List<LetterSequence> holdoverSequences = null;
        GroupOfShapes holdoverGroup = null;
        for (RowOfShapes row : paragraph.getRows()) {
            LOG.debug("Analysing row " + row.getIndex() + " (id=" + row.getId() + ")");
            for (GroupOfShapes group : row.getGroups()) {
                if (group.isSkip()) {
                    LOG.debug("Skipping group " + group.getIndex() + " (id=" + group.getId() + ")");
                    continue;
                }
                LOG.debug("Analysing group " + group.getIndex() + " (id=" + group.getId() + ")");
                int width = group.getRight() - group.getLeft() + 1;
                List<ShapeSequence> shapeSequences = null;
                if (boundaryDetector != null) {
                    shapeSequences = boundaryDetector.findBoundaries(group);
                } else {
                    // simply add this groups shape's
                    shapeSequences = new ArrayList<>();
                    ShapeSequence shapeSequence = new ShapeSequence();
                    for (Shape shape : group.getShapes()) shapeSequence.addShape(shape);
                    shapeSequences.add(shapeSequence);
                }
                // Perform a beam search to guess the most likely sequence
                // for this
                // word
                TreeMap<Integer, PriorityQueue<LetterSequence>> heaps = new TreeMap<>();
                // prime a starter heap with the n best shape boundary
                // analyses for
                // this group
                PriorityQueue<LetterSequence> starterHeap = new PriorityQueue<>(1);
                for (ShapeSequence shapeSequence : shapeSequences) {
                    LetterSequence emptySequence = new LetterSequence(shapeSequence, jochreSession);
                    starterHeap.add(emptySequence);
                }
                heaps.put(0, starterHeap);
                PriorityQueue<LetterSequence> finalHeap = null;
                while (heaps.size() > 0) {
                    Entry<Integer, PriorityQueue<LetterSequence>> heapEntry = heaps.pollFirstEntry();
                    if (LOG.isTraceEnabled())
                        LOG.trace("heap for index: " + heapEntry.getKey().intValue() + ", width: " + width);
                    if (heapEntry.getKey().intValue() == width) {
                        finalHeap = heapEntry.getValue();
                        break;
                    }
                    PriorityQueue<LetterSequence> previousHeap = heapEntry.getValue();
                    // limit the breadth to K
                    int maxSequences = previousHeap.size() > this.beamWidth ? this.beamWidth : previousHeap.size();
                    for (int j = 0; j < maxSequences; j++) {
                        LetterSequence history = previousHeap.poll();
                        ShapeInSequence shapeInSequence = history.getNextShape();
                        Shape shape = shapeInSequence.getShape();
                        if (LOG.isTraceEnabled()) {
                            LOG.trace("Sequence " + history + ", shape: " + shape);
                        }
                        LogUtils.logMemory(LOG);
                        int position = 0;
                        if (jochreSession.getLinguistics().isLeftToRight()) {
                            position = shape.getRight() - group.getLeft() + 1;
                        } else {
                            position = group.getRight() - shape.getLeft() + 1;
                        }
                        PriorityQueue<LetterSequence> heap = heaps.get(position);
                        if (heap == null) {
                            heap = new PriorityQueue<>();
                            heaps.put(position, heap);
                        }
                        letterGuesser.guessLetter(shapeInSequence, history);
                        // heap sort
                        for (Decision letterGuess : shape.getLetterGuesses()) {
                            // leave out very low probability outcomes
                            if (letterGuess.getProbability() > this.minOutcomeWeight) {
                                LetterSequence sequence = new LetterSequence(history);
                                sequence.getLetters().add(letterGuess.getOutcome());
                                sequence.addDecision(letterGuess);
                                heap.add(sequence);
                            }
                        // weight big enough to include
                        }
                    // next letter guess for this shape
                    }
                // next history in heap
                }
                // any more heaps?
                // find best sequence
                LetterSequence bestSequence = null;
                boolean isHoldover = false;
                List<LetterSequence> finalSequences = new ArrayList<>();
                for (int i = 0; i < this.beamWidth; i++) {
                    if (finalHeap.isEmpty())
                        break;
                    finalSequences.add(finalHeap.poll());
                }
                if (this.mostLikelyWordChooser == null) {
                    // most likely sequence is on top of the last heap
                    bestSequence = finalSequences.get(0);
                } else {
                    // get most likely sequence using lexicon
                    if (holdoverSequences != null) {
                        // we have a holdover from the previous row
                        // ending with a dash
                        bestSequence = this.mostLikelyWordChooser.chooseMostLikelyWord(finalSequences, holdoverSequences, this.beamWidth);
                    } else {
                        // check if this is the last group on the row
                        // and could end with
                        // a dash
                        boolean shouldBeHeldOver = false;
                        if (group.getIndex() == row.getGroups().size() - 1 && row.getIndex() < paragraph.getRows().size() - 1) {
                            for (LetterSequence letterSequence : finalSequences) {
                                if (letterSequence.toString().endsWith("-")) {
                                    shouldBeHeldOver = true;
                                    break;
                                }
                            }
                        }
                        if (shouldBeHeldOver) {
                            holdoverSequences = finalSequences;
                            holdoverGroup = group;
                            isHoldover = true;
                        } else {
                            // simplest case: no holdover
                            bestSequence = this.mostLikelyWordChooser.chooseMostLikelyWord(finalSequences, this.beamWidth);
                        }
                    }
                // have we holdover sequences?
                }
                if (!isHoldover) {
                    for (LetterGuessObserver observer : observers) {
                        observer.onBeamSearchEnd(bestSequence, finalSequences, holdoverSequences);
                    }
                }
                // assign letter
                if (!isHoldover) {
                    for (LetterGuessObserver observer : observers) {
                        observer.onStartSequence(bestSequence);
                    }
                    if (holdoverGroup == null) {
                        group.setBestLetterSequence(bestSequence);
                    } else {
                        // split bestSequence by group
                        List<LetterSequence> sequencesByGroup = bestSequence.splitByGroup();
                        for (LetterSequence sequenceByGroup : sequencesByGroup) {
                            if (sequenceByGroup.getGroups().get(0).equals(holdoverGroup))
                                holdoverGroup.setBestLetterSequence(sequenceByGroup);
                            else if (sequenceByGroup.getGroups().get(0).equals(group))
                                group.setBestLetterSequence(sequenceByGroup);
                        }
                        holdoverSequences = null;
                        holdoverGroup = null;
                    }
                    int i = 0;
                    for (ShapeInSequence shapeInSequence : bestSequence.getUnderlyingShapeSequence()) {
                        String bestOutcome = bestSequence.getLetters().get(i);
                        this.assignLetter(shapeInSequence, bestOutcome);
                        i++;
                    }
                    for (LetterGuessObserver observer : observers) {
                        observer.onGuessSequence(bestSequence);
                    }
                }
                this.shapeCount += group.getShapes().size();
                if (this.currentMonitor != null) {
                    double progress = (double) shapeCount / (double) totalShapeCount;
                    LOG.debug("progress: " + progress);
                    currentMonitor.setPercentComplete(progress);
                }
            }
        // next group
        }
    // next row
    }
    for (LetterGuessObserver observer : observers) {
        observer.onImageEnd();
    }
}

Example 18 with GroupOfShapes

use of com.joliciel.jochre.graphics.GroupOfShapes in project jochre by urieli.

the class JochreDocument method getXml.

/**
 * Returns an xml representation of this document as it currently stands, to
 * be used for correcting the text associated with this document.
 */
public void getXml(OutputStream outputStream) {
    try {
        XMLOutputFactory xmlOutputFactory = XMLOutputFactory.newFactory();
        XMLStreamWriter writer = xmlOutputFactory.createXMLStreamWriter(outputStream);
        writer.writeStartDocument("UTF-8", "1.0");
        writer.writeStartElement("doc");
        writer.writeAttribute("name", this.getName());
        writer.writeAttribute("fileName", this.getFileName());
        writer.writeAttribute("locale", this.getLocale().getLanguage());
        for (JochrePage page : this.getPages()) {
            writer.writeStartElement("page");
            writer.writeAttribute("index", "" + page.getIndex());
            for (JochreImage image : page.getImages()) {
                writer.writeStartElement("image");
                writer.writeAttribute("name", image.getName());
                writer.writeAttribute("index", "" + image.getIndex());
                for (Paragraph paragraph : image.getParagraphs()) {
                    writer.writeStartElement("paragraph");
                    writer.writeAttribute("index", "" + paragraph.getIndex());
                    StringBuffer sb = new StringBuffer();
                    for (RowOfShapes row : paragraph.getRows()) {
                        for (GroupOfShapes group : row.getGroups()) {
                            for (Shape shape : group.getShapes()) {
                                if (shape.getLetter() != null)
                                    sb.append(shape.getLetter());
                            }
                            sb.append(" ");
                        }
                        sb.append("\r\n");
                    }
                    writer.writeCData(sb.toString());
                    // paragraph
                    writer.writeEndElement();
                }
                // image
                writer.writeEndElement();
            }
            // page
            writer.writeEndElement();
        }
        // doc
        writer.writeEndElement();
        writer.writeEndDocument();
        writer.flush();
    } catch (XMLStreamException e) {
        throw new JochreException(e);
    }
}