Predicting sentence sentiment

predicting_sentence_sentiment

def predict_sentences(book, stop_words):
    #Break up book into sentences
    book_sentences = pd.DataFrame(book.split("."), columns = ['sentence'])
    
    #Clean sentences
    book_sentences['sentence'] = book_sentences['sentence'].\
        apply(lambda x: clean_labelled(x, stop_words))
    
    book_sentences = book_sentences[book_sentences['sentence'].\
                                    str.len() > 0] 
    
    book_sentences['classification'] = book_sentences['sentence'].\
                                                apply(nrc_sentence)
    
    #Adjust classifications
    book_sentences_adjusted = labelled_adjust_class(book_sentences)
    book_sentences_adjusted.index = np.arange(0, 
                    book_sentences_adjusted.shape[0])
    
    #Predict sentences' sentiment
    book_sentences_score = pd.concat([book_sentences_adjusted, pd.Series( \
            model.predict(book_sentences_adjusted.loc[:, 'anger':'trust']))], 
                                                             axis = 1)
    
    book_sentences_score = book_sentences_score.rename( \
                        columns = {0 : 'sentence_score'})
    
    return book_sentences_score
    
notes_sentece_predictions = predict_sentences(notes, stop_words_context)
crime_sentece_predictions = predict_sentences(crime, stop_words_context)
idiot_sentece_predictions = predict_sentences(idiot, stop_words_context)
possessed_sentece_predictions = predict_sentences(possessed, stop_words_context)
brothers_sentece_predictions = predict_sentences(brothers, stop_words_context)