How to use _get_args method in Molotov

Best Python code snippet using molotov_python

test_tra.py

Source:test_tra.py

...17    def setUpClass(cls):18        if cls is _AbstractTRATest:19            raise unittest.SkipTest('Skip abstract parent class')20        super(_AbstractTRATest, cls).setUpClass()21    def _get_args(self):22        # Add arguments for the Torch Ranker Agent to test23        # Override in child classes24        return dict(25            task='integration_tests:candidate',26            optimizer='adamax',27            learningrate=7e-3,28            batchsize=16,29            embedding_size=32,30            num_epochs=4,31        )32    def _get_threshold(self):33        # Accuracy threshold34        return 0.835    # test train inline cands36    @testing_utils.retry(ntries=3)37    def test_train_inline(self):38        args = self._get_args()39        args['candidates'] = 'inline'40        valid, test = testing_utils.train_model(args)41        threshold = self._get_threshold()42        self.assertGreaterEqual(valid['hits@1'], threshold)43    # test train batch cands44    @testing_utils.retry(ntries=3)45    def test_train_batch(self):46        args = self._get_args()47        args['candidates'] = 'batch'48        valid, test = testing_utils.train_model(args)49        threshold = self._get_threshold()50        self.assertGreaterEqual(valid['hits@1'], threshold)51    # test train fixed52    @testing_utils.retry(ntries=3)53    def test_train_fixed(self):54        args = self._get_args()55        args['candidates'] = 'fixed'56        args['encode_candidate_vecs'] = False57        valid, test = testing_utils.train_model(args)58        threshold = self._get_threshold()59        self.assertGreaterEqual(valid['hits@1'], threshold)60    # test train batch all cands61    @testing_utils.retry(ntries=3)62    def test_train_batch_all(self):63        args = self._get_args()64        args['candidates'] = 'batch-all-cands'65        valid, test = testing_utils.train_model(args)66        threshold = self._get_threshold()67        self.assertGreaterEqual(valid['hits@1'], threshold)68    # test eval inline ecands69    @testing_utils.retry(ntries=3)70    def test_eval_inline(self):71        args = self._get_args()72        args['eval_candidates'] = 'inline'73        valid, test = testing_utils.train_model(args)74        threshold = self._get_threshold()75        self.assertGreaterEqual(valid['hits@1'], threshold)76    # test eval batch ecands77    @testing_utils.retry(ntries=3)78    def test_eval_batch(self):79        args = self._get_args()80        args['eval_candidates'] = 'batch'81        valid, test = testing_utils.train_model(args)82        threshold = self._get_threshold()83        self.assertGreaterEqual(valid['hits@1'], threshold)84    # test eval fixed ecands85    @testing_utils.retry(ntries=3)86    def test_eval_fixed(self):87        args = self._get_args()88        args['eval_candidates'] = 'fixed'89        args['encode_candidate_vecs'] = True90        args['ignore_bad_candidates'] = True91        valid, test = testing_utils.train_model(args)92        # none of the train candidates appear in evaluation, so should have93        # zero accuracy: this tests whether the fixed candidates were built94        # properly (i.e., only using candidates from the train set)95        self.assertEqual(valid['hits@1'], 0)96        # now try again with a fixed candidate file that includes all possible97        # candidates98        teacher = CandidateTeacher({'datatype': 'train'})99        all_cands = teacher.train + teacher.val + teacher.test100        all_cands_str = '\n'.join([' '.join(x) for x in all_cands])101        with testing_utils.tempdir() as tmpdir:102            tmp_cands_file = os.path.join(tmpdir, 'all_cands.text')103            with open(tmp_cands_file, 'w') as f:104                f.write(all_cands_str)105            args['fixed_candidates_path'] = tmp_cands_file106            args['encode_candidate_vecs'] = False  # don't encode before training107            args['ignore_bad_candidates'] = False108            args['num_epochs'] = 4109            valid, test = testing_utils.train_model(args)110            self.assertGreaterEqual(valid['hits@100'], 0.1)111    # test eval vocab ecands112    @testing_utils.retry(ntries=3)113    def test_eval_vocab(self):114        args = self._get_args()115        args['eval_candidates'] = 'vocab'116        args['encode_candidate_vecs'] = True117        valid, test = testing_utils.train_model(args)118        # accuracy should be zero, none of the vocab candidates should be the119        # correct label120        self.assertEqual(valid['hits@100'], 0)121class TestTransformerRanker(_AbstractTRATest):122    def _get_args(self):123        args = super()._get_args()124        new_args = dict(125            model='transformer/ranker',126            n_layers=1,127            n_heads=4,128            ffn_size=64,129            gradient_clip=0.5,130        )131        for k, v in new_args.items():132            args[k] = v133        return args134class TestMemNN(_AbstractTRATest):135    def _get_args(self):136        args = super()._get_args()137        args['model'] = 'memnn'138        return args139    def _get_threshold(self):140        # this is a slightly worse model, so we expect it to perform worse141        return 0.5142class TestPolyRanker(_AbstractTRATest):143    def _get_args(self):144        args = super()._get_args()145        new_args = dict(146            model='transformer/polyencoder',147            n_layers=1,148            n_heads=4,149            ffn_size=64,150            gradient_clip=0.5,151        )152        for k, v in new_args.items():153            args[k] = v154        return args155    def _get_threshold(self):156        return 0.6157    def test_eval_fixed_label_not_in_cands(self):158        # test where cands during eval do not contain test label159        args = self._get_args()160        args[161            'model'162        ] = 'parlai.agents.transformer.polyencoder:IRFriendlyPolyencoderAgent'163        args['eval_candidates'] = 'fixed'164        teacher = CandidateTeacher({'datatype': 'train'})165        all_cands = teacher.train + teacher.val + teacher.test166        train_val_cands = teacher.train + teacher.val167        all_cands_str = '\n'.join([' '.join(x) for x in all_cands])168        train_val_cands_str = '\n'.join([' '.join(x) for x in train_val_cands])169        with testing_utils.tempdir() as tmpdir:170            tmp_cands_file = os.path.join(tmpdir, 'all_cands.text')171            with open(tmp_cands_file, 'w') as f:172                f.write(all_cands_str)173            tmp_train_val_cands_file = os.path.join(tmpdir, 'train_val_cands.text')...

test_run.py

Source:test_run.py

...4import numpy as np5import os6import pandas as pd7import pytest8def _get_args(infile: str, k: list, outdir: str):9    args = RUN_DEFAULTS.copy()10    args['outdir'] = outdir11    args['k'] = k12    args["infile"] = infile13    return args14def test_init(tmpdir):15    # incorrect parameters16    with pytest.raises(AttributeError):17        SumoRun()18    fname = os.path.join(tmpdir, "data.npz")19    outdir = os.path.join(tmpdir, "outdir")20    samples = 1021    sample_labels = ['sample_{}'.format(i) for i in range(samples)]22    args = _get_args(fname, [2], outdir)23    # no input file24    with pytest.raises(FileNotFoundError):25        SumoRun(**args)26    save_arrays_to_npz({'0': np.random.random((samples, samples)),27                        '1': np.random.random((samples, samples)),28                        'samples': np.array(sample_labels)}, fname)29    # supervised sumo no labels file30    labels_fname = os.path.join(tmpdir, "labels.tsv")31    labels = np.array([sample_labels[0:2] + sample_labels[7:9], [1, 1, 2, 2]]).T32    labels_df = pd.DataFrame(data=labels, columns=['sample', 'label'])33    args = _get_args(fname, [2], outdir)34    args['labels'] = labels_fname35    with pytest.raises(FileNotFoundError):36        SumoRun(**args)37    labels_df.to_csv(labels_fname, sep="\t")38    args['sparsity'] = [10]39    args['n'] = 10  # makes test run quicker40    SumoRun(**args)41    # incorrect number of repetitions42    args['n'] = -143    with pytest.raises(ValueError):44        SumoRun(**args)45    # incorrect number of threads46    args = _get_args(fname, [2], outdir)47    args['t'] = -148    with pytest.raises(ValueError):49        SumoRun(**args)50    # incorrect outdir51    args = _get_args(fname, [2], fname)52    with pytest.raises(NotADirectoryError):53        SumoRun(**args)54    # incorrect k55    args = _get_args(fname, [2, 3, 4], outdir)56    with pytest.raises(ValueError):57        SumoRun(**args)58    args = _get_args(fname, [2], outdir)59    SumoRun(**args)60    args = _get_args(fname, [2, 5], outdir)61    SumoRun(**args)62    # incorrect k range63    args = _get_args(fname, [5, 2], outdir)64    with pytest.raises(ValueError):65        SumoRun(**args)66    # incorrect subsample argument67    args = _get_args(fname, [2], outdir)68    args['subsample'] = -0.169    with pytest.raises(ValueError):70        SumoRun(**args)71    args['subsample'] = 0.972    with pytest.raises(ValueError):73        SumoRun(**args)74def test_run(tmpdir):75    fname = os.path.join(tmpdir, "data.npz")76    outdir = os.path.join(tmpdir, "outdir")77    samples = 1078    a0 = np.random.random((samples, samples))79    a0 = (a0 * a0.T) / 280    a1 = np.random.random((samples, samples))81    a1 = (a1 * a1.T) / 282    sample_labels = ['sample_{}'.format(i) for i in range(samples)]83    args = _get_args(fname, [2], outdir)84    # no sample names85    save_arrays_to_npz({'0': a0, '1': a1}, fname)86    with pytest.raises(ValueError):87        sr = SumoRun(**args)88        sr.run()89    # incorrect sample names90    save_arrays_to_npz({'0': a0, '1': a1, 'samples': np.array(sample_labels[1:])}, fname)91    with pytest.raises(ValueError):92        sr = SumoRun(**args)93        sr.run()94    # incorrect adjacency matrices95    save_arrays_to_npz({'samples': np.array(sample_labels)}, fname)96    with pytest.raises(ValueError):97        sr = SumoRun(**args)98        sr.run()99    # incorrect value of h_init100    save_arrays_to_npz({'0': a0, '1': a1, 'samples': np.array(sample_labels)}, fname)101    args = _get_args(fname, [2], outdir)102    args['h_init'] = -1103    with pytest.raises(ValueError):104        sr = SumoRun(**args)105        sr.run()106    args['h_init'] = 3107    with pytest.raises(ValueError):108        sr = SumoRun(**args)109        sr.run()110    # supervised sumo111    labels_fname = os.path.join(tmpdir, "labels.tsv")112    labels = np.array([sample_labels[0:2] + sample_labels[7:9], [1, 1, 2, 2]]).T113    labels_df = pd.DataFrame(data=labels, columns=['sample', 'label'])114    labels_df.to_csv(labels_fname, sep="\t")115    args = _get_args(fname, [2], outdir)116    args['labels'] = labels_fname117    args['sparsity'] = [10]118    args['n'] = 10  # makes test run quicker119    sr = SumoRun(**args)120    sr.run()121    # unsupervised sumo122    args = _get_args(fname, [2], outdir)123    args['sparsity'] = [10]124    args['n'] = 10  # makes test run quicker125    sr = SumoRun(**args)126    sr.run()127    assert all([os.path.exists(os.path.join(outdir, x)) for x in ['k2', 'plots',128                                                                  os.path.join('plots', 'consensus_k2.png'),...

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.