"""! @brief Unit-tests for G-Means algorithm. @authors Andrei Novikov (pyclustering@yandex.ru) @date 2014-2020 @copyright BSD-3-Clause """ import unittest # Generate images without having a window appear. import matplotlib matplotlib.use('Agg') from pyclustering.cluster.tests.gmeans_templates import gmeans_test_template from pyclustering.cluster.gmeans import gmeans from pyclustering.utils import read_sample, distance_metric, type_metric from pyclustering.samples.definitions import SIMPLE_SAMPLES, FCPS_SAMPLES, SIMPLE_ANSWERS class testable_gmeans(gmeans): @staticmethod def get_data_projection(data, vector): return gmeans._project_data(data, vector) class gmeans_unit_test(unittest.TestCase): def test_data_projection(self): data = [[1, 1], [2, 2], [3, 3]] projection = testable_gmeans.get_data_projection(data, [2, 2]) self.assertEqual([0.5, 1.0, 1.5], projection.tolist()) projection = testable_gmeans.get_data_projection(data, [1, 1]) self.assertEqual([1.0, 2.0, 3.0], projection.tolist()) def test_clustering_sample_01(self): gmeans_test_template().clustering(SIMPLE_SAMPLES.SAMPLE_SIMPLE1, SIMPLE_ANSWERS.ANSWER_SIMPLE1, 1, False) def test_clustering_sample_01_kmax_correct(self): gmeans_test_template().clustering(SIMPLE_SAMPLES.SAMPLE_SIMPLE1, SIMPLE_ANSWERS.ANSWER_SIMPLE1, 1, False, k_max=2) def test_clustering_sample_01_kmax_1(self): gmeans_test_template().clustering(SIMPLE_SAMPLES.SAMPLE_SIMPLE1, [10], 1, False, k_max=1) def test_clustering_sample_01_kmax_10(self): gmeans_test_template().clustering(SIMPLE_SAMPLES.SAMPLE_SIMPLE1, SIMPLE_ANSWERS.ANSWER_SIMPLE1, 1, False, k_max=10) def test_clustering_sample_02(self): gmeans_test_template().clustering(SIMPLE_SAMPLES.SAMPLE_SIMPLE2, SIMPLE_ANSWERS.ANSWER_SIMPLE2, 1, False) def test_clustering_sample_02_kmax_correct(self): gmeans_test_template().clustering(SIMPLE_SAMPLES.SAMPLE_SIMPLE2, SIMPLE_ANSWERS.ANSWER_SIMPLE2, 1, False, k_max=3) def test_clustering_sample_02_kmax_1(self): gmeans_test_template().clustering(SIMPLE_SAMPLES.SAMPLE_SIMPLE2, [23], 1, False, k_max=1) def test_clustering_sample_02_kmax_2(self): gmeans_test_template().clustering(SIMPLE_SAMPLES.SAMPLE_SIMPLE2, [8, 15], 1, False, k_max=2) def test_clustering_sample_03(self): gmeans_test_template().clustering(SIMPLE_SAMPLES.SAMPLE_SIMPLE3, SIMPLE_ANSWERS.ANSWER_SIMPLE3, 1, False, random_state=1000) def test_clustering_sample_03_kmax_1(self): gmeans_test_template().clustering(SIMPLE_SAMPLES.SAMPLE_SIMPLE3, [60], 1, False, k_max=1, random_state=1000) def test_clustering_sample_05(self): gmeans_test_template().clustering(SIMPLE_SAMPLES.SAMPLE_SIMPLE5, SIMPLE_ANSWERS.ANSWER_SIMPLE5, 1, False) def test_clustering_sample_06(self): gmeans_test_template().clustering(SIMPLE_SAMPLES.SAMPLE_SIMPLE6, SIMPLE_ANSWERS.ANSWER_SIMPLE6, 1, False) def test_clustering_sample_07(self): gmeans_test_template().clustering(SIMPLE_SAMPLES.SAMPLE_SIMPLE7, SIMPLE_ANSWERS.ANSWER_SIMPLE7, 1, False) def test_clustering_sample_08(self): gmeans_test_template().clustering(SIMPLE_SAMPLES.SAMPLE_SIMPLE8, SIMPLE_ANSWERS.ANSWER_SIMPLE8, 1, False) def test_clustering_sample_09(self): gmeans_test_template().clustering(SIMPLE_SAMPLES.SAMPLE_SIMPLE9, SIMPLE_ANSWERS.ANSWER_SIMPLE9, 1, False) def test_clustering_sample_10(self): gmeans_test_template().clustering(SIMPLE_SAMPLES.SAMPLE_SIMPLE10, SIMPLE_ANSWERS.ANSWER_SIMPLE10, 1, False) def test_clustering_sample_11(self): gmeans_test_template().clustering(SIMPLE_SAMPLES.SAMPLE_SIMPLE11, SIMPLE_ANSWERS.ANSWER_SIMPLE11, 1, False) def test_clustering_sample_12(self): gmeans_test_template().clustering(SIMPLE_SAMPLES.SAMPLE_SIMPLE12, SIMPLE_ANSWERS.ANSWER_SIMPLE12, 1, False) def test_clustering_sample_13(self): gmeans_test_template().clustering(SIMPLE_SAMPLES.SAMPLE_SIMPLE13, SIMPLE_ANSWERS.ANSWER_SIMPLE13, 1, False) def test_clustering_sample_15(self): gmeans_test_template().clustering(SIMPLE_SAMPLES.SAMPLE_SIMPLE15, SIMPLE_ANSWERS.ANSWER_SIMPLE15, 1, False, random_state=1200) def test_clustering_hepta_kmax_01(self): gmeans_test_template().clustering(FCPS_SAMPLES.SAMPLE_HEPTA, 1, 1, False, k_max=1) def test_clustering_hepta_kmax_02(self): gmeans_test_template().clustering(FCPS_SAMPLES.SAMPLE_HEPTA, 2, 1, False, k_max=2, random_state=1000) def test_clustering_hepta_kmax_03(self): gmeans_test_template().clustering(FCPS_SAMPLES.SAMPLE_HEPTA, 3, 1, False, k_max=3, random_state=1000) def test_clustering_hepta_kmax_04(self): gmeans_test_template().clustering(FCPS_SAMPLES.SAMPLE_HEPTA, 4, 1, False, k_max=4, random_state=1000) def test_clustering_hepta_kmax_05(self): gmeans_test_template().clustering(FCPS_SAMPLES.SAMPLE_HEPTA, 5, 1, False, k_max=5, random_state=1000) def test_clustering_hepta_kmax_06(self): gmeans_test_template().clustering(FCPS_SAMPLES.SAMPLE_HEPTA, 6, 1, False, k_max=6, random_state=1000) def test_clustering_hepta_kmax_07(self): gmeans_test_template().clustering(FCPS_SAMPLES.SAMPLE_HEPTA, 7, 1, False, k_max=7, random_state=1000) def test_clustering_hepta_kmax_08(self): gmeans_test_template().clustering(FCPS_SAMPLES.SAMPLE_HEPTA, 7, 1, False, k_max=8, random_state=1000) def test_clustering_hepta_kmax_09(self): gmeans_test_template().clustering(FCPS_SAMPLES.SAMPLE_HEPTA, 7, 1, False, k_max=9, random_state=1000) def test_clustering_hepta_kmax_10(self): gmeans_test_template().clustering(FCPS_SAMPLES.SAMPLE_HEPTA, 7, 1, False, k_max=10, random_state=1000) def test_incorrect_data(self): self.assertRaises(ValueError, gmeans, [], 1, False, tolerance=0.05, repeat=3) def test_incorrect_k_init(self): self.assertRaises(ValueError, gmeans, [[0], [1], [2]], -1, False, tolerance=0.05, repeat=3) self.assertRaises(ValueError, gmeans, [[0], [1], [2]], 0, False, tolerance=0.05, repeat=3) def test_incorrect_tolerance(self): self.assertRaises(ValueError, gmeans, [[0], [1], [2]], 3, False, tolerance=0, repeat=3) def test_incorrect_repeat(self): self.assertRaises(ValueError, gmeans, [[0], [1], [2]], 3, False, tolerance=1, repeat=0) def test_incorrect_kmax_zero(self): self.assertRaises(ValueError, gmeans, [[0], [1], [2]], 3, False, tolerance=1, repeat=0, kmax=0) def test_incorrect_kmax_negative(self): self.assertRaises(ValueError, gmeans, [[0], [1], [2]], 3, False, tolerance=1, repeat=0, kmax=-2) def test_kmax_less_than_kinit(self): self.assertRaises(ValueError, gmeans, [[0], [1], [2]], 3, False, tolerance=1, repeat=0, kmax=2) def test_predict_without_process(self): self.assertEqual([], gmeans([[0], [1]]).predict([0])) def template_predict(self, path, amount, points, ccore): metric = distance_metric(type_metric.EUCLIDEAN) sample = read_sample(path) gmeans_instance = gmeans(sample, amount, ccore).process() centers = gmeans_instance.get_centers() closest_clusters = gmeans_instance.predict(points) self.assertEqual(len(points), len(closest_clusters)) for i in range(len(points)): cluster_index = closest_clusters[i] distance = metric(centers[cluster_index], points[i]) for center_index in range(len(centers)): if center_index != cluster_index: other_distance = metric(centers[center_index], points[i]) self.assertLessEqual(distance, other_distance) def test_predict_one_point(self): self.template_predict(SIMPLE_SAMPLES.SAMPLE_SIMPLE3, 4, [[0.3, 0.2]], False) self.template_predict(SIMPLE_SAMPLES.SAMPLE_SIMPLE3, 4, [[4.1, 1.1]], False) self.template_predict(SIMPLE_SAMPLES.SAMPLE_SIMPLE3, 4, [[2.1, 1.9]], False) self.template_predict(SIMPLE_SAMPLES.SAMPLE_SIMPLE3, 4, [[2.1, 4.1]], False) def test_predict_two_points(self): self.template_predict(SIMPLE_SAMPLES.SAMPLE_SIMPLE3, 4, [[0.3, 0.2], [2.1, 1.9]], False) self.template_predict(SIMPLE_SAMPLES.SAMPLE_SIMPLE3, 4, [[2.1, 4.1], [2.1, 1.9]], False)