I've got an Algorithm class whose responsibility is to find if a given word is in a list of words.

As part of doing that, the algorithm first has to lowercase the words, remove punctuation, and remove any stopwords. Then, finally, the algorithm tries to find the given word in the cleaned list via the find_word method.

As you can see in the code, I have split everything in small methods to avoid having one big method. However, I am not sure what's the proper way to call those methods. I am calling all them inside the find_word method one after the other, but that smells like a bad design to me. How do I do this the right way?

import string
from nltk.corpus import stopwords
stopwords = stopwords.words('english')

class Algorithm:

    def __init__(self, words: list):
        self.words = words

    def _lower(self):
        # Lowercase all words of the list
        self.words = [word.lower() for word in self.words]

    def _remove_punctuation(self):
        # Remove punctuation from all words of the list
        punct_dict = dict((ord(punct), None) for punct in string.punctuation)  # creates dict of {33:None, 34:None, etc}
        self.words = [words.translate(punct_dict) for words in

    def _remove_stopwords(self):
        # Remove stopwords from list such as in, at, who, etc.
        self.words = [word for word in self.words if word not in stopwords]

    def find_word(self, user_word):
        # Find if the user given word is in the cleaned list
        if user_word in self.words:
            return user_word

class Chatbot:

    def answer(self, user_word):
        algorithm = Algorithm(['Rabbit', 'Horse', 'turtle'])
        return algorithm.find_word(user_word)

chatbot = Chatbot()

1 Answer 1


The only obviously bad thing in your design is that if you call find_word multiple times on the same Algorithm instance, you are duplicating the preparation work on the wordlist.

That can be remedied by calling _lower, _remove_punctuation and _remove_stopwords from __init__.

Apart from that, you might want to do some preprocessing on the user_word as well, so that find_word('hORSE') would also work, and maybe even find_word('!horse!').

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.