KaneAI - World’s First E2E Software Testing Agent.Book a Demo Join TestMu Conference

How to use test_alphabets method in autotest

Best Python code snippet using autotest_python

preprocess_omni.py

Source:preprocess_omni.py

1import argparse2import numpy as np3import os4import pickle5from scipy.io import loadmat6parser = argparse.ArgumentParser()7parser.add_argument('--data-dir', type=str, default="/home/gigi/ns_data/omniglot_ns")8def _load():9    # load data10    file = os.path.join(args.data_dir, 'chardata.mat')11    data = loadmat(file)12    # data is in train/test split so read separately13    train_images = data['data'].astype(np.float32).T14    train_alphabets = np.argmax(data['target'].astype(np.float32).T, axis=1)15    train_characters = data['targetchar'].astype(np.float32)16    test_images = data['testdata'].astype(np.float32).T17    test_alphabets = np.argmax(data['testtarget'].astype(np.float32).T, axis=1)18    test_characters = data['testtargetchar'].astype(np.float32)19    # combine train and test data20    images = np.concatenate([train_images, test_images], axis=0)21    alphabets = np.concatenate([train_alphabets, test_alphabets], axis=0)22    characters = np.concatenate([np.ravel(train_characters),23                                 np.ravel(test_characters)], axis=0)24    data = (images, alphabets, characters)25    return data26# def load():27#     # load data28#     file = os.path.join(args.data_dir, 'chardata.mat')29#     data = loadmat(file)30#     # data is in train/test split so read separately31#     train_images = data['data'].astype(np.float32).T32#     train_alphabets = np.argmax(data['target'].astype(np.float32).T, axis=1)33#     train_characters = data['targetchar'].astype(np.float32)34#     tr_images = np.concatenate([train_images], axis=0)35#     tr_alphabets = np.concatenate([train_alphabets], axis=0)36#     tr_characters = np.concatenate([np.ravel(train_characters)], axis=0)37#     test_images = data['testdata'].astype(np.float32).T38#     test_alphabets = np.argmax(data['testtarget'].astype(np.float32).T, axis=1)39#     test_characters = data['testtargetchar'].astype(np.float32)40#     # combine train and test data41#     ts_images = np.concatenate([test_images], axis=0)42#     ts_alphabets = np.concatenate([test_alphabets], axis=0)43#     ts_characters = np.concatenate([np.ravel(test_characters)], axis=0)44    45#     tr_data = (tr_images, tr_alphabets, tr_characters)46#     ts_data = (ts_images, ts_alphabets, ts_characters)47#     return tr_data, ts_data48# def modify(data):49#     # We don't care about alphabets, so combine all alphabets50#     # into a single character ID.51#     # First collect all unique (alphabet, character) pairs.52#     images, alphabets, characters = data53#     unique_alphabet_character_pairs = list(set(zip(alphabets, characters)))54#     # Now assign each pair an ID55#     ids = np.asarray([unique_alphabet_character_pairs.index((alphabet, character))56#                       for (alphabet, character) in zip(alphabets, characters)])57#     # Now split into train(1200)/val(323)/test(100) by character58#     # train_images = images[ids < 1200]59#     # train_labels = ids[ids < 1200]60#     # # val_images = images[(1200 <= ids) * (ids < 1523)]61#     # val_labels = ids[(1200 <= ids) * (ids < 1523)]62#     # test_images = images[1523 <= ids]63#     # test_labels = ids[1523 <= ids]64#     # split_data = (train_images, train_labels, 65#     #               val_images, val_labels, 66#     #               test_images, test_labels)67#     return (images, ids)68# def _modify(data):69#     # We don't care about alphabets, so combine all alphabets70#     # into a single character ID.71#     # First collect all unique (alphabet, character) pairs.72#     images, alphabets, characters = data73#     unique_alphabet_character_pairs = list(set(zip(alphabets, characters)))74#     # Now assign each pair an ID75#     ids = np.asarray([unique_alphabet_character_pairs.index((alphabet, character))76#                       for (alphabet, character) in zip(alphabets, characters)])77#     print(ids.shape)78#     print(images.shape)79#     # Now split into train(1200)/val(323)/test(100) by character80#     train_images = images[ids < 1200]81#     train_labels = ids[ids < 1200]82#     test_images = images[1200 <= ids]83#     test_labels = ids[1200 <= ids]84#     #val_images = images[(1200 <= ids) * (ids < 1523)]85#     #val_labels = ids[(1200 <= ids) * (ids < 1523)]86#     #test_images = images[1523 <= ids]87#     #test_labels = ids[1523 <= ids]88#     print(train_images.shape, test_images.shape)89#     print(train_labels.shape, test_labels.shape)90#     split_data = (train_images, train_labels, 91#                   test_images, test_labels)92#     return split_data93def _modify(data):94    # We don't care about alphabets, so combine all alphabets95    # into a single character ID.96    # First collect all unique (alphabet, character) pairs.97    images, alphabets, characters = data98    unique_alphabet_character_pairs = list(set(zip(alphabets, characters)))99    # Now assign each pair an ID100    ids = np.asarray([unique_alphabet_character_pairs.index((alphabet, character))101                      for (alphabet, character) in zip(alphabets, characters)])102    print(ids.shape)103    print(images.shape)104    # Now split into train(1000)/val(200)/test(460) by character105    train_images = images[ids < 1000]106    train_labels = ids[ids < 1000]107    108    val_images = images[(1000 <= ids) * (ids < 1200)]109    val_labels = ids[(1000 <= ids) * (ids < 1200)]110    test_images = images[1200 <= ids]111    test_labels = ids[1200 <= ids]112    113    print(train_images.shape, val_images.shape, test_images.shape)114    print(train_labels.shape, val_labels.shape, test_labels.shape)115    split_data = (train_images, train_labels,116                  val_images, val_labels, 117                  test_images, test_labels)118    return split_data119# def main():120#     tr, ts = load()121#     tr_data = modify(tr)122#     ts_data = modify(ts)123#     tr_img, tr_lbl = tr_data124#     ts_img, ts_lbl = ts_data125#     print(tr_img.shape, tr_lbl.shape, ts_img.shape, ts_lbl.shape)126#     data = (tr_img, tr_lbl, ts_img, ts_lbl)127#     #save(data)128def save(data):129    savepath = os.path.join(args.data_dir, 'omni_train_val_test.pkl')130    with open(savepath, 'wb') as file:131        pickle.dump(data, file)132def _main():133    data = _load()134    modified_data = _modify(data)135    save(modified_data)136if __name__ == '__main__':137    args = parser.parse_args()138    assert (args.data_dir is not None) and (os.path.isdir(args.data_dir))...

omnicreate.py

Source:omnicreate.py

1import argparse2import numpy as np3import os4import pickle5from scipy.io import loadmat6parser = argparse.ArgumentParser()7parser.add_argument('--data-dir', required=True, type=str, default=None)8args = parser.parse_args()9assert (args.data_dir is not None) and (os.path.isdir(args.data_dir))10def load():11    # load data12    file = os.path.join(args.data_dir, 'chardata.mat')13    data = loadmat(file)14    # data is in train/test split so read separately15    train_images = data['data'].astype(np.float32).T16    train_alphabets = np.argmax(data['target'].astype(np.float32).T, axis=1)17    train_characters = data['targetchar'].astype(np.float32)18    test_images = data['testdata'].astype(np.float32).T19    test_alphabets = np.argmax(data['testtarget'].astype(np.float32).T, axis=1)20    test_characters = data['testtargetchar'].astype(np.float32)21    # combine train and test data22    images = np.concatenate([train_images, test_images], axis=0)23    alphabets = np.concatenate([train_alphabets, test_alphabets], axis=0)24    characters = np.concatenate([np.ravel(train_characters),25                                 np.ravel(test_characters)], axis=0)26    data = (images, alphabets, characters)27    return data28def modify(data):29    # We don't care about alphabets, so combine all alphabets30    # into a single character ID.31    # First collect all unique (alphabet, character) pairs.32    images, alphabets, characters = data33    unique_alphabet_character_pairs = list(set(zip(alphabets, characters)))34    # Now assign each pair an ID35    ids = np.asarray([unique_alphabet_character_pairs.index((alphabet, character))36                      for (alphabet, character) in zip(alphabets, characters)])37    # Now split into train(1200)/val(323)/test(100) by character38    train_images = images[ids < 1200]39    train_labels = ids[ids < 1200]40    val_images = images[(1200 <= ids) * (ids < 1523)]41    val_labels = ids[(1200 <= ids) * (ids < 1523)]42    test_images = images[1523 <= ids]43    test_labels = ids[1523 <= ids]44    split_data = (train_images, train_labels, val_images,45                  val_labels, test_images, test_labels)46    return split_data47def save(data):48    savepath = os.path.join(args.data_dir, 'train_val_test_split.pkl')49    with open(savepath, 'wb') as file:50        pickle.dump(data, file)51def main():52    data = load()53    modified_data = modify(data)54    save(modified_data)55if __name__ == '__main__':...

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.