private Optional init()

in modules/ml-ext/ml/src/main/java/org/apache/ignite/ml/clustering/gmm/GmmTrainer.java [419:472]
45 lines of code
7 McCabe index (conditional complexity)

    private Optional<GmmModel> init(Dataset<EmptyContext, GmmPartitionData> dataset) {
        int cntOfTries = 0;

        while (true) {
            try {
                if (initialMeans == null) {
                    List<Vector> randomMeansSets = Stream.of(dataset.compute(
                        selectNRandomXsMapper(countOfComponents),
                        GmmTrainer::selectNRandomXsReducer))
                        .flatMap(Stream::of)
                        .sorted(Comparator.comparingDouble(Vector::getLengthSquared)).collect(Collectors.toList());

                    Collections.shuffle(randomMeansSets, environment.randomNumbersGenerator());

                    A.ensure(
                        randomMeansSets.size() >= countOfComponents,
                        "There is not enough data in dataset for select N random means"
                    );

                    initialMeans = randomMeansSets.subList(0, countOfComponents)
                        .toArray(new Vector[countOfComponents]);
                }

                dataset.compute(data -> GmmPartitionData.estimateLikelihoodClusters(data, initialMeans));

                List<Matrix> initCovs = CovarianceMatricesAggregator.computeCovariances(
                    dataset,
                    VectorUtils.fill(1. / countOfComponents, countOfComponents),
                    initialMeans
                );

                if (initCovs.isEmpty())
                    return Optional.empty();

                List<MultivariateGaussianDistribution> distributions = new ArrayList<>();
                for (int i = 0; i < countOfComponents; i++)
                    distributions.add(new MultivariateGaussianDistribution(initialMeans[i], initCovs.get(i)));

                return Optional.of(new GmmModel(
                    VectorUtils.of(DoubleStream.generate(() -> 1. / countOfComponents).limit(countOfComponents).toArray()),
                    distributions
                ));
            }
            catch (SingularMatrixException | IllegalArgumentException e) {
                String msg = "Cannot construct non-singular covariance matrix by data. " +
                    "Try to select other initial means or other model trainer [number of tries = " + cntOfTries + "]";
                environment.logger().log(MLLogger.VerboseLevel.HIGH, msg);
                cntOfTries += 1;
                initialMeans = null;
                if (cntOfTries >= maxCountOfInitTries)
                    throw new RuntimeException(msg, e);
            }
        }
    }