Java Code Examples for com.google.common.collect.Multiset#size()

private static void assertBalancing(BucketBalancer balancer, int expectedMoves)
{
    int actualMoves = balancer.balance();
    assertEquals(actualMoves, expectedMoves);

    // check that number of buckets per node is within bounds
    ClusterState clusterState = balancer.fetchClusterState();
    for (Distribution distribution : clusterState.getDistributionAssignments().keySet()) {
        Multiset<String> allocationCounts = HashMultiset.create();
        clusterState.getDistributionAssignments().get(distribution).stream()
                .map(BucketAssignment::getNodeIdentifier)
                .forEach(allocationCounts::add);

        double bucketsPerNode = (1.0 * allocationCounts.size()) / clusterState.getActiveNodes().size();
        for (String node : allocationCounts) {
            assertGreaterThanOrEqual(allocationCounts.count(node), (int) Math.floor(bucketsPerNode), node + " has fewer buckets than expected");
            assertLessThanOrEqual(allocationCounts.count(node), (int) Math.ceil(bucketsPerNode), node + " has more buckets than expected");
        }
    }

    // check stability
    assertEquals(balancer.balance(), 0);
}
 

public ScorerContext context(List<String> words) {
    Multiset<String> counter = TreeMultiset.create();
    counter.addAll(words);

    int word_dim = kb.getEmbeddingsSize();
    // word_vecs is the concatenation of all word vectors of the word list
    float[] word_vecs = new float[counter.size() * word_dim];
    IntArrayList word_counts = new IntArrayList();
    int n_words = 0;

    for(Multiset.Entry<String> entry : counter.entrySet()) {
        short[] vector = kb.getWordEmbeddings(entry.getElement());
        if (vector != null) {
            word_counts.add(entry.getCount());
            for (int i=0; i<kb.getEmbeddingsSize(); i++) {
                word_vecs[n_words * word_dim + i] = vector[i];
            }
            n_words += 1;
        }
    }
    word_counts.trim();

    return create_context(word_vecs, word_counts.elements());
}
 

private static void addEntryForWord(final Multiset<Category> countForCategory,
		final Map<String, Collection<Category>> result, final String word) {
	final List<Entry<Category>> cats = new ArrayList<>();
	for (final Entry<Category> catToCount : countForCategory.entrySet()) {
		cats.add(catToCount);
	}
	final int totalSize = countForCategory.size();
	final int minSize = Math.floorDiv(totalSize, 1000);
	Collections.sort(cats, comparator);
	final List<Category> cats2 = new ArrayList<>();

	for (final Entry<Category> entry : cats) {
		if (entry.getCount() >= minSize) {
			cats2.add(entry.getElement());
		}
	}

	result.put(word, cats2);
}
 

NeuralNetworkTrainer(NeuralNetworkConfig config, Multiset<String> vocab, Map<String, HuffmanNode> huffmanNodes, TrainingProgressListener listener) {
	this.config = config;
	this.huffmanNodes = huffmanNodes;
	this.listener = listener;
	this.vocabSize = huffmanNodes.size();
	this.numTrainedTokens = vocab.size();
	this.layer1_size = config.layerSize;
	this.window = config.windowSize;
	
	this.actualWordCount = new AtomicInteger();
	this.alpha = config.initialLearningRate;
	
	this.syn0 = new double[vocabSize][layer1_size];
	this.syn1 = new double[vocabSize][layer1_size];
	this.syn1neg = new double[vocabSize][layer1_size];
	this.table = new int[TABLE_SIZE];
	
	initializeSyn0();
	initializeUnigramTable();
}
 

private static PartialCreditTriplet partialCompare(LogicalExpression gold,
		LogicalExpression label) {
	final Multiset<Pair<? extends LogicalExpression, ? extends LogicalExpression>> goldPairs = GetPredConstPairs
			.of(gold);
	final Multiset<Pair<? extends LogicalExpression, ? extends LogicalExpression>> labelPairs;
	if (label == null) {
		labelPairs = HashMultiset.create();
	} else {
		labelPairs = GetPredConstPairs.of(label);
	}

	// The "intersection" of the gold and label pair sets = the number of
	// matches
	final Multiset<Pair<? extends LogicalExpression, ? extends LogicalExpression>> intersection = HashMultiset
			.create();

	for (final Entry<Pair<? extends LogicalExpression, ? extends LogicalExpression>> entry : goldPairs
			.entrySet()) {
		intersection.setCount(
				entry.getElement(),
				Math.min(entry.getCount(),
						labelPairs.count(entry.getElement())));
	}

	return new PartialCreditTriplet(goldPairs.size(), labelPairs.size(),
			intersection.size());
}
 

@Override
public boolean isFrequent(final Category category, final int argumentNumber, final SRLLabel label) {
	if (label == SRLFrame.NONE) {
		return true;
	}
	final Multiset<SRLLabel> countForCategory = categoryToArgumentToSRLs.get(category.withoutAnnotation(),
			argumentNumber);
	return countForCategory != null && countForCategory.size() >= minSlotSRL
			&& countForCategory.count(label) >= minSlotRole;
}
 

@Override
public double getMLProbability(final A element, final B given) {
	if (table.containsKey(given)) {
		final Multiset<A> elements = table.get(given);
		return ((double) elements.count(element)) / elements.size();
	} else {
		return 1;
	}
}
 

@Override
public double getMLProbability(final A element, final B given) {
	if (table.containsKey(given)) {
		final Multiset<A> elements = table.get(given);
		return ((double) elements.count(element)) / elements.size();
	} else {
		return 1;
	}
}
 

@Override
public int size() {
	final Multiset<E> multiset = get();
	return (multiset == null) ? EMPTY_MULTISET.size() : multiset.size();
}
 

private List<byte[]> getUnitigs(final LocusIteratorByState libs) {
    final List<StringBuilder> unitigBuilders = new ArrayList<>();
    int lastCoveredLocus = Integer.MIN_VALUE;
    while (libs.hasNext()) {
        final ReadPileup pileup = libs.next().getBasePileup();
        if (pileup.isEmpty()) {
            continue;
        }

        // begin new unitig if this pileup isn't contiguous with the last
        final int currentLocus = pileup.getLocation().getStart();
        if (currentLocus != lastCoveredLocus + 1) {
            unitigBuilders.add(new StringBuilder());
        }
        lastCoveredLocus = currentLocus;
        final StringBuilder currentUnitigBuilder = unitigBuilders.get(unitigBuilders.size() - 1);

        // add no bases (deletion) or consensus bases.
        final int[] baseCounts = pileup.getBaseCounts();
        final int deletionCount = (int) Utils.stream(pileup).filter(PileupElement::isDeletion).count();
        if (deletionCount < pileup.size() / 2) {
            final byte consensusBase = BaseUtils.baseIndexToSimpleBase(MathUtils.maxElementIndex(baseCounts));
            currentUnitigBuilder.append((char) consensusBase);

            // in addition to consensus base, add inserted bases if needed
            final Multiset<String> insertedBases = Utils.stream(pileup)
                    .map(PileupElement::getBasesOfImmediatelyFollowingInsertion)
                    .filter(s -> s != null)
                    .collect(Collectors.toCollection(HashMultiset::create));

            if (insertedBases.size() > pileup.size() / 2) {
                final String consensusInsertion = Multisets.copyHighestCountFirst(insertedBases).entrySet().iterator().next().getElement();
                currentUnitigBuilder.append(consensusInsertion);
            }
        }
    }

    return unitigBuilders.stream()
            .map(builder -> builder.toString().getBytes())
            .filter(unitig -> unitig.length > MIN_UNITIG_LENGTH)
            .collect(Collectors.toList());
}
 

private void check(Multiset<String> counts, double p, String s) {
    double n = counts.size();
    assertEquals(p, counts.count(s) / n, Math.sqrt(n * p * (n - p)));
}