How to use test_set_target method in avocado

Best Python code snippet using avocado_python

test_optimizer.py

Source:test_optimizer.py

1import numpy as np2import unittest3from icet.fitting import Optimizer4class TestOptimizer(unittest.TestCase):5    """Unittest class for Optimizer."""6    def __init__(self, *args, **kwargs):7        super().__init__(*args, **kwargs)8        self.n_rows = 2009        self.n_cols = 5010        self.tol = 1.0 / self.n_rows11        # set up dummy linear problem data12        self.A = np.random.normal(0, 1, (self.n_rows, self.n_cols))13        self.x = np.random.normal(0, 5, (self.n_cols, ))14        self.noise = np.random.normal(0, 0.1, (self.n_rows, ))15        self.y = np.dot(self.A, self.x) + self.noise16    def shortDescription(self):17        """Prevents unittest from printing docstring in test cases."""18        return None19    def test_get_rows_via_sizes(self):20        """Tests _get_rows_via_sizes functionality."""21        opt = Optimizer((self.A, self.y))22        # test with only train_size defined23        train_size, test_size = int(0.8 * self.n_rows), None24        train_set, test_set = opt._get_rows_via_sizes(train_size, test_size)25        self.assertEqual(train_size, len(train_set))26        self.assertEqual(self.n_rows - train_size, len(test_set))27        # test with only test_size defined28        train_size, test_size = None, int(0.8 * self.n_rows)29        train_set, test_set = opt._get_rows_via_sizes(train_size, test_size)30        self.assertEqual(test_size, len(test_set))31        self.assertEqual(self.n_rows - test_size, len(train_set))32        # test with both defined33        train_size, test_size = int(0.8 * self.n_rows), int(0.15 * self.n_rows)34        train_set, test_set = opt._get_rows_via_sizes(train_size, test_size)35        self.assertEqual(train_size, len(train_set))36        self.assertEqual(test_size, len(test_set))37        # test with fractions38        train_size, test_size = 0.7, 0.239        train_set, test_set = opt._get_rows_via_sizes(train_size, test_size)40        self.assertLess(abs(train_size*self.n_rows - len(train_set)), self.tol)41        self.assertLess(abs(test_size*self.n_rows - len(test_set)), self.tol)42        # test edge case with full training set43        test_size = None44        for train_size in [1.0, self.n_rows]:45            train_set, test_set = opt._get_rows_via_sizes(46                train_size, test_size)47            self.assertEqual(len(train_set), self.n_rows)48            self.assertIsNone(test_set)49        # test invalid sizes50        with self.assertRaises(ValueError):51            train_size, test_size = None, 1.052            opt._get_rows_via_sizes(train_size, test_size)53        with self.assertRaises(ValueError):54            train_size, test_size = None, None55            opt._get_rows_via_sizes(train_size, test_size)56    def test_get_rows_from_indices(self):57        """Tests _get_rows_from_indices."""58        opt = Optimizer((self.A, self.y))59        all_rows = np.arange(self.n_rows)60        train_size = int(0.8 * self.n_rows)61        train_set_target = np.random.choice(62            all_rows, train_size, replace=False)63        test_set_target = sorted(np.setdiff1d(all_rows, train_set_target))64        # specify only train_set, test set should default to remaining rows65        train_set, test_set = opt._get_rows_from_indices(66            train_set_target, None)67        self.assertSequenceEqual(sorted(train_set_target), sorted(train_set))68        self.assertSequenceEqual(sorted(test_set_target), sorted(test_set))69        # specify only test_set, train set should default to remaining rows70        train_set, test_set = opt._get_rows_from_indices(None, test_set_target)71        self.assertSequenceEqual(sorted(train_set_target), sorted(train_set))72        self.assertSequenceEqual(sorted(test_set_target), sorted(test_set))73        # specify partial sets meaning not all rows are used74        train_set_target = np.delete(train_set_target, [0, 1, 2])75        test_set_target = np.delete(test_set_target, [0, 1, 2])76        train_set, test_set = opt._get_rows_from_indices(77            train_set_target, test_set_target)78        self.assertSequenceEqual(sorted(train_set_target), sorted(train_set))79        self.assertSequenceEqual(sorted(test_set_target), sorted(test_set))80        # test invalid input81        with self.assertRaises(ValueError):82            opt._get_rows_from_indices(None, None)83    def test_setup_rows(self):84        """85        Tests _setup_rows.86        Simply test that function raise when no training data available87        """88        opt = Optimizer((self.A, self.y))89        # no training data from train_size90        with self.assertRaises(ValueError):91            train_size, test_size = 0, 0.592            opt._setup_rows(train_size, test_size, None, None)93        # no training data from train_set94        with self.assertRaises(ValueError):95            train_set, test_set = [], np.arange(0, int(0.5*self.n_rows))96            opt._setup_rows(None, None, train_set, test_set)97        # overlapping indices in train_set and test_set98        with self.assertRaises(ValueError):99            train_set, test_set = [1, 2, 3, 4, 5], [5, 6, 7, 8, 9, 10]100            opt._setup_rows(None, None, train_set, test_set)101    def test_train(self):102        """Tests train."""103        # with test set104        train_size = 0.75105        opt = Optimizer((self.A, self.y), train_size=train_size)106        self.assertIsNone(opt._rmse_train)107        self.assertIsNone(opt._rmse_test)108        self.assertIsNone(opt.train_scatter_data)109        self.assertIsNone(opt.test_scatter_data)110        opt.train()111        self.assertIsNotNone(opt._rmse_train)112        self.assertIsNotNone(opt._rmse_test)113        self.assertIsNotNone(opt.train_scatter_data)114        self.assertIsNotNone(opt.test_scatter_data)115        # without testing116        train_size = 1.0117        opt = Optimizer((self.A, self.y), train_size=train_size)118        self.assertIsNone(opt._rmse_train)119        self.assertIsNone(opt._rmse_test)120        self.assertIsNone(opt.train_scatter_data)121        self.assertIsNone(opt.test_scatter_data)122        opt.train()123        self.assertIsNotNone(opt._rmse_train)124        self.assertIsNone(opt._rmse_test)125        self.assertIsNotNone(opt.train_scatter_data)126        self.assertIsNone(opt.test_scatter_data)127    def test_summary_property(self):128        """Tests summary property."""129        # without having trained130        opt = Optimizer((self.A, self.y))131        self.assertIsInstance(opt.summary, dict)132        # with having trained133        opt.train()134        self.assertIsInstance(opt.summary, dict)135        self.assertIn('rmse_train', opt.summary.keys())136        self.assertIn('rmse_test', opt.summary.keys())137    def test_repr(self):138        """Tests repr dunder."""139        opt = Optimizer((self.A, self.y))140        self.assertIsInstance(repr(opt), str)141    def test_size_properties(self):142        """Tests the properties in regards to training/test sets and sizes."""143        # test without test_set144        train_set = np.arange(0, self.n_rows)145        opt = Optimizer((self.A, self.y), train_set=train_set)146        self.assertSequenceEqual(opt.train_set.tolist(), train_set.tolist())147        self.assertEqual(len(train_set), opt.train_size)148        self.assertEqual(1.0, opt.train_fraction)149        self.assertIsNone(opt.test_set)150        self.assertEqual(opt.test_size, 0)151        self.assertEqual(opt.test_fraction, 0)152        # test with test set153        test_set = np.arange(int(0.7 * self.n_rows), int(0.8 * self.n_rows))154        opt = Optimizer((self.A, self.y), test_set=test_set)155        self.assertSequenceEqual(test_set.tolist(), opt.test_set.tolist())156        self.assertEqual(opt.test_size, len(test_set))157        self.assertAlmostEqual(opt.test_fraction, len(test_set) / self.n_rows)158    def test_zero_error_with_least_square_fit(self):159        """ Test that the error is zero if training without noise and with160        least-squares. """161        # set up dummy linear problem data162        for standardize in [True, False]:163            y = np.dot(self.A, self.x)164            opt = Optimizer((self.A, y), fit_method='least-squares', standardize=standardize)165            opt.train()166            self.assertAlmostEqual(opt.rmse_train, 0.0)167            self.assertAlmostEqual(opt.rmse_test, 0.0)168            self.assertAlmostEqual(np.abs(self.x - opt.parameters).max(), 0)169if __name__ == '__main__':...

week11.py

Source:week11.py

1from __future__ import division2import hashlib3import json4from sklearn.ensemble import RandomForestClassifier as rfc5directory = "data/"6def bag_of_words(lines):7    uniq = set()8    for line in lines:9        # Iterate over all words10        for word in line[0].split():11            # Put words not already found into dictionary12            uniq.add(word)13    uniq_dict = {}14    for j, word in enumerate(uniq):15        uniq_dict[word] = j16    uniq_len = len(uniq)17    # Use word dictionary to create bag-of-words18    bag = []19    for line in lines:20        vec = [0] * uniq_len21        for word in line[0].split():22            vec[uniq_dict[word]] += 123        bag.append((vec, (1 if 'earn' in line[1] else 0)))24    return bag25def bag_of_words_feature_hashed(lines, N):26    uniq = set()27    for line in lines:28        # Iterate over all words29        for word in line[0].split():30            # Put words not already found into dictionary31            uniq.add(word)32    uniq_dict = {}33    for j, word in enumerate(uniq):34        uniq_dict[word] = j35    # Use word dictionary to create bag-of-words36    bag = []37    for line in lines:38        vec = [0] * N39        for word in line[0].split():40            h = int(hashlib.md5(word).hexdigest(), 16) % N41            vec[h] += 142        bag.append((vec, (1 if 'earn' in line[1] else 0)))43    return bag44if __name__ == "__main__":45    articles = []46    for i in range(21):47        f = open(directory + "reuters-0" + (str(i) if i >= 10 else "0" + str(i)) + ".json", 'r')48        articles += json.loads(f.read())49        f.close()50    articles = reduce(lambda x, y: x + ([y] if ('topics' in y.keys())51                                               and ('body' in y.keys())52                                               and (len(y['topics']) > 0)53                                               and (len(y['body']) > 0)54                                            else []), articles, [])55    lines = map(lambda x: (x['body'].lower().encode('ascii', errors='ignore'), x['topics']), articles)56    print "Calcualting bag of words."57    bow = bag_of_words(lines)58    print "Amount of lines: " + str(len(bow))59    print "Amount of words: " + str(len(bow[0][0]))60    training_set = bow[:int(round(len(bow)*0.8))]61    training_set_data = [row[0] for row in training_set]62    training_set_target = [row[1] for row in training_set]63    test_set = bow[-int(round(len(bow)*0.2)):]64    test_set_data = [row[0] for row in test_set]65    test_set_target = [row[1] for row in test_set]66    classifier = rfc(n_estimators=50)67    classifier.fit(training_set_data, training_set_target)68    predictions = classifier.predict(test_set_data)69    accuracy = []70    for i, prediction in enumerate(predictions):71        accuracy += [test_set_target[i] == prediction]72    accuracy = (accuracy.count(True) / len(test_set_target)) * 10073    print "Accuracy of classifier: " + str(accuracy) + "%"74    print "\nCalculating bag of words using feature hashing."75    bow = bag_of_words_feature_hashed(lines, 1000)76    print "Amount of lines: " + str(len(bow))77    print "Amount of words: " + str(len(bow[0][0]))78    training_set = bow[:int(round(len(bow)*0.8))]79    training_set_data = [row[0] for row in training_set]80    training_set_target = [row[1] for row in training_set]81    test_set = bow[-int(round(len(bow)*0.2)):]82    test_set_data = [row[0] for row in test_set]83    test_set_target = [row[1] for row in test_set]84    classifier = rfc(n_estimators=50)85    classifier.fit(training_set_data, training_set_target)86    predictions = classifier.predict(test_set_data)87    accuracy = []88    for i, prediction in enumerate(predictions):89        accuracy += [test_set_target[i] == prediction]90    accuracy = (accuracy.count(True) / len(test_set_target)) * 100...

evaluate_digit_classifier.py

Source:evaluate_digit_classifier.py

1import numpy as np2from sklearn import datasets3# initialize test set4def load_digits_and_test_set(test_start=0, test_count=1000):5    global digits6    global t_start, t_count, t_end7    global test_set_images, test_set_target8    digits = datasets.load_digits()9    t_images = len(digits.images)10    t_start = test_start11    t_count = test_count12    t_end = t_start + t_count13    if (t_images < t_end):14        t_end = t_images15    test_set_images = digits.images[t_start:t_end]16    test_set_target = digits.target[t_start:t_end]17    return len(test_set_images)18# test classifier over test set19def test_digit_classifier(classifier):20    digits = datasets.load_digits()21    correct = 022    print(f'test classifier for {t_count} images: {t_start} to {t_end-1}.')23    for img, target in zip(test_set_images, test_set_target):24        v = np.matrix.flatten(img) / 15.25        output = classifier(v)26        answer = list(output).index(max(output))27        if answer == target:28            correct += 129    return (correct/t_count)30# calculate total cost for classifier over test set31def y_vec(digit):32    return np.array([1 if i == digit else 0 for i in range(0,10)])33def cost_one(classifier,x,i):34    return sum([(classifier(x)[j] - y_vec(i)[j])**2 for j in range(10)])35def calculate_total_cost(classifier):36    digits = datasets.load_digits()37    print(f'calculate total cost for {t_count} images: {t_start} to {t_end-1}.')38    x = np.array([np.matrix.flatten(img) for img in test_set_images]) / 15.039    y = test_set_target...

Automation Testing Tutorials

Learn to execute automation testing from scratch with LambdaTest Learning Hub. Right from setting up the prerequisites to run your first automation test, to following best practices and diving deeper into advanced test scenarios. LambdaTest Learning Hubs compile a list of step-by-step guides to help you be proficient with different test automation frameworks i.e. Selenium, Cypress, TestNG etc.