def test_evaluate_average_accuracy_svm():
"""Test evaluate average accuracy for support vector machine"""
classifier = ReviewClassifier('svm')
average = classifier.evaluate_average_accuracy('test-reviews.csv')
assert average > 0
assert average < 100
os.remove('output.log')
def test_create_trained_model_rf():
"""Test create trained model for random forest"""
classifier = ReviewClassifier('rf')
train_data, train_target = classifier.build_dataset('test-reviews.csv')
model = classifier.create_trained_model(train_data=train_data,
train_target=train_target)
assert model
def test_evaluate_accuracy_nn():
"""Test evaluate_accuracy for nn model"""
classifer = ReviewClassifier('nn')
classifer.load_model(tar_file='test_nn_model.tar')
score = classifer.evaluate_accuracy('test-reviews.csv')
assert score > 0
os.remove('output.log')
def test_evaluate_average_accuracy_nb():
"""Test evaluate average accuracy for naive bayes"""
classifier = ReviewClassifier('nb')
average = classifier.evaluate_average_accuracy('test-reviews.csv')
assert average > 0
assert average < 100
os.remove('output.log')
def test_evaluate_average_accuracy_rf():
"""Test evaluate average accuracy for random forest"""
classifier = ReviewClassifier('rf')
average = classifier.evaluate_average_accuracy('test-reviews.csv')
assert average > 0
assert average < 100
os.remove('output.log')
def test_evaluate_accuracy_rf():
"""Test evaluate_accuracy for rf model"""
classifer = ReviewClassifier('rf')
classifer.load_model(pickle_file='test_rf_model.pickle')
score = classifer.evaluate_accuracy('test-reviews.csv')
assert score > 0
os.remove('output.log')
def test_load_svm_model():
"""Test load svm model from pickle file"""
classifier = ReviewClassifier('svm')
classifier.load_model(pickle_file='test_svm_model.pickle')
assert classifier.model
assert classifier.vectorizer
assert classifier.encoder
def test_create_trained_model_nn():
"""Test create trained model for neural network"""
classifier = ReviewClassifier('nn')
train_data, train_target = classifier.build_dataset('test-reviews.csv')
model = classifier.create_trained_model(train_data=train_data,
train_target=train_target)
assert model
def test_load_nn_model():
"""Test load nn model from tar file"""
classifier = ReviewClassifier('nn')
classifier.load_model(tar_file='test_nn_model.tar')
assert classifier.model
assert classifier.vectorizer
assert classifier.encoder
def test_split_data_target():
classifier = ReviewClassifier('nn')
dataset = classifier.create_dataset('test-reviews.csv')
train_data, train_target = classifier.split_data_target(dataset)
assert len(train_data) == len(train_target)
assert train_data[0][0] in [0, 1]
assert train_target[0] in [0, 1]
def test_create_trained_model_svm():
"""Test create trained model for support vector machine"""
classifier = ReviewClassifier('svm')
train_data, train_target = classifier.build_dataset('test-reviews.csv')
model = classifier.create_trained_model(train_data=train_data,
train_target=train_target)
assert model
def test_evaluate_average_accuracy_nn():
"""Test evaluate average accuracy for neural network"""
classifier = ReviewClassifier('nn')
average = classifier.evaluate_average_accuracy('test-reviews.csv')
assert average > 0
assert average < 100
os.remove('output.log')
def test_save_nn_model():
"""Test save nn model"""
classifier = ReviewClassifier('nn')
classifier.train('test-reviews.csv')
classifier.save_model()
assert os.path.exists('trained_nn_model.tar')
os.remove('trained_nn_model.tar')
def test_save_svm_model():
"""Test save svm model"""
classifier = ReviewClassifier('svm')
classifier.train('test-reviews.csv')
classifier.save_model()
assert os.path.exists('trained_svm_model.pickle')
os.remove('trained_svm_model.pickle')
def test_classify_rf():
"""Test classify using rf"""
classifer = ReviewClassifier('rf')
classifer.load_model(pickle_file='test_rf_model.pickle')
classifer.classify('classified_comments.txt',
comments_filename='test-reviews.csv')
assert os.path.exists('classified_comments.txt')
os.remove('classified_comments.txt')
def test_classify_nb_from_csv():
"""Test classify using nb model from csv file"""
classifer = ReviewClassifier('nb')
classifer.load_model(pickle_file='test_nb_model.pickle')
classifer.classify('classified_comments.txt',
comments_filename='test-reviews.csv')
assert os.path.exists('classified_comments.txt')
os.remove('classified_comments.txt')
def test_classify_from_text_file():
"""Test classify comments text file"""
classifer = ReviewClassifier('nb')
classifer.load_model(pickle_file='test_nb_model.pickle')
classifer.classify('classified_comments.txt',
comments_text_file='neutral.txt')
assert os.path.exists('classified_comments.txt')
os.remove('classified_comments.txt')
def main():
""" Main function.
"""
# Review sentiment classifier
review_classifier = ReviewClassifier(
'nb') # Try 'nb', 'dt', 'rf', and 'nn'
review_classifier.train('citalopram-reviews.csv')
review_classifier.save_model()
def test_classify_nn():
"""Test classify using nn"""
classifer = ReviewClassifier('nn')
classifer.load_model(tar_file='test_nn_model.tar')
classifer.classify('classified_comments.txt',
comments_filename='test-reviews.csv')
assert os.path.exists('classified_comments.txt')
os.remove('classified_comments.txt')
def test_load_model_nb_no_file():
"""Test load naive bayes model without model"""
classifier = ReviewClassifier('nb')
with pytest.raises(Exception):
classifier.load_model()
def test_create_dataset():
classifier = ReviewClassifier('nb')
dataset = classifier.create_dataset('test-reviews.csv')
assert list(dataset[0][0].values())[0]
assert dataset[0][1] == 'pos'
assert dataset[-1][1] == 'neg'
def test_train_svm():
"""Test train for support vector machine"""
classifier = ReviewClassifier('svm')
classifier.train('test-reviews.csv')
assert classifier.model is not None
def test_train_nb():
"""Test train for naive bayes"""
classifier = ReviewClassifier('nb')
classifier.train('test-reviews.csv')
assert classifier.model is not None
def test_train_rf():
"""Test train for random forest"""
classifier = ReviewClassifier('rf')
classifier.train('test-reviews.csv')
assert classifier.model is not None
def test_train_nn():
"""Test train for neural network"""
classifier = ReviewClassifier('nn')
classifier.train('test-reviews.csv')
assert classifier.model is not None
def test_load_model_nn_no_file():
"""Test load nn model without model"""
classifier = ReviewClassifier('nn')
with pytest.raises(Exception):
classifier.load_model()
def test_load_nb_model():
"""Test load nb model from pickle file"""
classifier = ReviewClassifier('nb')
classifier.load_model(pickle_file='test_nb_model.pickle')
assert classifier.model
def test_create_trained_model_nb():
"""Test create train model for naive bayes"""
classifier = ReviewClassifier('nb')
dataset = classifier.create_dataset('test-reviews.csv')
model = classifier.create_trained_model(dataset=dataset)
assert model
def test_build_dataset():
classifier = ReviewClassifier('nn')
train_data, train_target = classifier.build_dataset('test-reviews.csv')
assert len(train_data) == len(train_target)
assert train_data[0][0] in [0, 1]
assert train_target[0] in [0, 1]