from sklearn.metrics import accuracy_score
if __name__ == '__main__':
pre = Preprocessing()
df_input = pre.data
train_data = pre.train_data
test_data = pre.test_data
Y_train = pre.Y_train
Y_test = pre.Y_test
lng = Language(df_input)
print("Preparing the language data")
train_tokens = train_data['title'].apply(util.get_tokens)
lng_data_train = lng.get_encoded_data(train_tokens)
test_tokens = test_data['title'].apply(util.get_tokens)
lng_data_test = lng.get_encoded_data(test_tokens)
language_model = lng.lng_model
print("training the language model (bi-lstm), this might take some time")
language_model.fit(lng_data_train, Y_train, verbose=1, validation_split=0.2, nb_epoch=5)
## printing precision_recall- language modality
Y_pred = language_model.predict(lng_data_test, verbose=1)
y_pred = np.array([np.argmax(pred) for pred in Y_pred])
print("******************language modality scores(unimodal)*******************************")
print(' Classification Report:\n', classification_report(Y_test, y_pred), '\n')
image_mod = Image_modality()
