def test_from_tsv_equals_to_tsv(self):
test_args = args.copy()
test_args['out-file'] = tmpdir + '/tmp.tsv'
converter.fiddle(dict_to_obj(test_args))
expected = pd.read_csv(test_args['in-file'])
created = pd.read_csv(test_args['out-file'])
self.assertTrue(expected.equals(created))
def test_from_asc_equals_to_asc(self):
test_args = args.copy()
test_args['in-file'] = test_asc_filepath
test_args['out-file'] = tmpdir + '/tmp.asc'
test_args['in-format'] = 'asc'
test_args['out-format'] = 'asc'
converter.fiddle(dict_to_obj(test_args))
self.assertTrue(
filecmp.cmp(test_args['in-file'], test_args['out-file']))
def test_train_test_split_creates_train_and_test_files(self):
test_args = args.copy()
test_args['test-rate'] = 0.2
test_args = dict_to_obj(test_args)
original_data_len = len(pd.read_csv(test_args.in_file))
test_data_expected_len = int(original_data_len * test_args.test_rate)
train_data_expected_len = original_data_len - test_data_expected_len
converter.fiddle(test_args)
self.assertEqual(train_data_expected_len,
len(pd.read_csv(test_args.out_file + '.train')))
self.assertEqual(test_data_expected_len,
len(pd.read_csv(test_args.out_file + '.test')))
def test_from_csv_to_yudelson_bkt(self):
test_args = args.copy()
test_args['in-file'] = test_simple_csv_filepath
test_args['out-file'] = tmpdir + '/tmp.tsv'
test_args['exercise-col'] = 'exercise_id'
test_args['in-format'] = 'csv'
test_args['out-format'] = 'yudelson-bkt'
converter.fiddle(dict_to_obj(test_args))
expected = pd.read_csv(test_args['in-file'])
created = pd.read_csv(test_args['out-file'], sep='\t', header=None)
expected.correct = -(expected.correct - 1) + 1
expected = expected[['correct', 'user_id', 'exercise_id', 'skill_id']]
expected.user_id = 'student' + expected.user_id.astype(str)
expected.exercise_id = 'exercise' + expected.exercise_id.astype(str)
expected.skill_id = 'skill' + expected.skill_id.astype(str)
print(expected)
print(created)
self.assertTrue(np.alltrue(expected.values == created.values))