"""! @brief Unit-tests for genetic clustering algorithm. @authors Andrei Novikov, Aleksey Kukushkin (pyclustering@yandex.ru) @date 2014-2020 @copyright BSD-3-Clause """ import unittest import inspect import numpy # Generate images without having a window appear. import matplotlib matplotlib.use('Agg') from pyclustering.samples.definitions import SIMPLE_SAMPLES from pyclustering.cluster.ga import genetic_algorithm, ga_observer, ga_visualizer from pyclustering.cluster.ga_maths import ga_math from pyclustering.utils import read_sample class GeneticAlgorithmClusteringUnitTest(unittest.TestCase): def runGeneticAlgorithm(self, test_case_name, data, count_chromosomes, count_clusters, count_populations, count_mutations_gen, result_should_be): _, best_ff, = genetic_algorithm(data=data, count_clusters=count_clusters, chromosome_count=count_chromosomes, population_count=count_populations, count_mutation_gens=count_mutations_gen, random_state=1000).process() # Check result self.assertEqual(best_ff, result_should_be) def test1CenterClustering(self): data = [[0, 0], [0, 2]] self.runGeneticAlgorithm(test_case_name=inspect.stack()[0][3], data=data, count_chromosomes=10, count_clusters=1, count_populations=10, count_mutations_gen=1, result_should_be=2.0) def test1Center4DataClustering(self): data = [[0, 0], [0, 2], [2, 0], [2, 2]] self.runGeneticAlgorithm(test_case_name=inspect.stack()[0][3], data=data, count_chromosomes=10, count_clusters=1, count_populations=10, count_mutations_gen=1, result_should_be=8.0) def test2Center8DataClustering(self): data = [[0, 0], [0, 2], [2, 0], [2, 2]] data.extend([[6, 0], [6, 2], [8, 0], [8, 2]]) self.runGeneticAlgorithm(test_case_name=inspect.stack()[0][3], data=data, count_chromosomes=50, count_clusters=2, count_populations=50, count_mutations_gen=1, result_should_be=16.0) def test4Center16DataClustering(self): data = [] data.extend([[0, 0], [1, 0], [0, 1], [1, 1]]) data.extend([[5, 0], [6, 0], [5, 1], [6, 1]]) data.extend([[0, 5], [1, 5], [0, 6], [1, 6]]) data.extend([[4, 4], [7, 4], [4, 7], [7, 7]]) self.runGeneticAlgorithm(test_case_name=inspect.stack()[0][3], data=data, count_chromosomes=20, count_clusters=4, count_populations=100, count_mutations_gen=1, result_should_be=24.0) def templateDataClustering(self, sample_path, amount_clusters, chromosome_count, population_count, count_mutation_gens, coeff_mutation_count, expected_clusters_sizes, **kwargs): scale_points = kwargs.get('scale_points', None) sample = numpy.array(read_sample(sample_path)) if scale_points is not None: sample = sample * scale_points ga_instance = genetic_algorithm(sample, amount_clusters, chromosome_count, population_count, count_mutations_gen=count_mutation_gens, coeff_mutation_count=coeff_mutation_count, **kwargs) ga_instance.process() clusters = ga_instance.get_clusters() obtained_cluster_sizes = [len(cluster) for cluster in clusters] self.assertEqual(len(sample), sum(obtained_cluster_sizes)) if expected_clusters_sizes is not None: obtained_cluster_sizes.sort() expected_clusters_sizes.sort() self.assertEqual(obtained_cluster_sizes, expected_clusters_sizes) def testClusteringTwoDimensionalData(self): self.templateDataClustering(SIMPLE_SAMPLES.SAMPLE_SIMPLE1, 2, 20, 20, 2, 0.25, [5, 5], random_state=1000) def testClusteringTwoDimensionalDataWrongAllocation(self): self.templateDataClustering(SIMPLE_SAMPLES.SAMPLE_SIMPLE1, 1, 20, 20, 2, 0.25, [10], random_state=1000) def testClusteringNonNormalizedValues(self): self.assertRaises(ValueError, self.templateDataClustering, SIMPLE_SAMPLES.SAMPLE_SIMPLE1, 2, 20, 20, 2, 0.25, [5, 5], random_state=1000, scale_points=1000) def testClusteringSimple02(self): self.templateDataClustering(SIMPLE_SAMPLES.SAMPLE_SIMPLE2, 3, 20, 40, 2, 0.25, [5, 8, 10], random_state=1000) def testClusteringSimple09(self): self.templateDataClustering(SIMPLE_SAMPLES.SAMPLE_SIMPLE9, 2, 20, 20, 2, 0.25, [10, 20], random_state=1000) def testClusteringOneDimensionalData(self): self.templateDataClustering(SIMPLE_SAMPLES.SAMPLE_SIMPLE7, 2, 20, 20, 2, 0.25, [10, 10], random_state=1000) def testClusteringOneDimensionalDataWrongAllocation(self): self.templateDataClustering(SIMPLE_SAMPLES.SAMPLE_SIMPLE7, 1, 20, 20, 2, 0.25, [20], random_state=1000) def testClusteringThreeDimensionalData(self): self.templateDataClustering(SIMPLE_SAMPLES.SAMPLE_SIMPLE11, 2, 20, 20, 2, 0.25, [10, 10], random_state=1000) def testClusteringThreeDimensionalDataWrongAllocation(self): self.templateDataClustering(SIMPLE_SAMPLES.SAMPLE_SIMPLE11, 1, 20, 20, 2, 0.25, [20], random_state=1000) def testTwoClustersTotallySimilarObjects(self): self.templateDataClustering(SIMPLE_SAMPLES.SAMPLE_SIMPLE12, 2, 20, 20, 2, 0.25, None, random_state=1000) def testFiveClustersTotallySimilarObjects(self): self.templateDataClustering(SIMPLE_SAMPLES.SAMPLE_SIMPLE12, 5, 20, 20, 2, 0.25, None, random_state=1000) def testTenClustersTotallySimilarObjects(self): self.templateDataClustering(SIMPLE_SAMPLES.SAMPLE_SIMPLE12, 10, 20, 20, 2, 0.25, None, random_state=1000) def templateTestObserverCollecting(self, amount_clusters, iterations, global_optimum, local_optimum, average, **kwargs): observer_instance = ga_observer(global_optimum, local_optimum, average) self.assertEqual(0, len(observer_instance)) sample = read_sample(SIMPLE_SAMPLES.SAMPLE_SIMPLE1) ga_instance = genetic_algorithm(sample, amount_clusters, 20, iterations, count_mutation_gens=2, coeff_mutation_count=0.25, observer=observer_instance, **kwargs) ga_instance.process() self.assertEqual(observer_instance, ga_instance.get_observer()) expected_length = 0 if global_optimum is True: expected_length = iterations + 1 self.assertEqual(expected_length, len(observer_instance)) self.assertEqual(expected_length, len(observer_instance.get_global_best()['chromosome'])) self.assertEqual(expected_length, len(observer_instance.get_global_best()['fitness_function'])) expected_length = 0 if local_optimum is True: expected_length = iterations + 1 self.assertEqual(expected_length, len(observer_instance)) self.assertEqual(expected_length, len(observer_instance.get_population_best()['chromosome'])) self.assertEqual(expected_length, len(observer_instance.get_population_best()['fitness_function'])) expected_length = 0 if average is True: expected_length = iterations + 1 self.assertEqual(expected_length, len(observer_instance)) self.assertEqual(expected_length, len(observer_instance.get_mean_fitness_function())) if global_optimum is True: clusters = ga_math.get_clusters_representation(observer_instance.get_global_best()['chromosome'][-1]) self.assertEqual(amount_clusters, len(clusters)) return sample, observer_instance def testObserveGlobalOptimum(self): self.templateTestObserverCollecting(2, 10, True, False, False, random_state=1000) def testObserveLocalOptimum(self): self.templateTestObserverCollecting(2, 11, False, True, False, random_state=1000) def testObserveAverage(self): self.templateTestObserverCollecting(2, 12, False, False, True, random_state=1000) def testObserveAllParameters(self): self.templateTestObserverCollecting(2, 9, True, True, True, random_state=1000) def testObserveNoCollecting(self): self.templateTestObserverCollecting(2, 9, False, False, False, random_state=1000) def testObserveParameterCombinations(self): self.templateTestObserverCollecting(3, 10, True, True, False, random_state=1000) self.templateTestObserverCollecting(4, 10, True, False, True, random_state=1000) self.templateTestObserverCollecting(1, 10, False, True, True, random_state=1000) def testNoFailureVisualizationApi(self): sample, observer = self.templateTestObserverCollecting(2, 10, True, True, True, random_state=1000) ga_visualizer.show_evolution(observer) ga_visualizer.show_clusters(sample, observer) ga_visualizer.animate_cluster_allocation(sample, observer) def testNoFailureShowEvolution(self): _, observer = self.templateTestObserverCollecting(2, 10, True, True, True, random_state=1000) ga_visualizer.show_evolution(observer, 2, 5) ga_visualizer.show_evolution(observer, 2, len(observer)) ga_visualizer.show_evolution(observer, 2, len(observer), display=False) def testNoneObserver(self): sample = read_sample(SIMPLE_SAMPLES.SAMPLE_SIMPLE1) ga_instance = genetic_algorithm(sample, 2, 20, 20, count_mutation_gens=2, coeff_mutation_count=0.25, observer=None) ga_instance.process() self.assertIsNone(ga_instance.get_observer()) def test_incorrect_data(self): self.assertRaises(ValueError, genetic_algorithm, [], 1, 2, 2) def test_incorrect_amount_clusters(self): self.assertRaises(ValueError, genetic_algorithm, [[0], [1], [2]], 0, 2, 2)