WordleSolve.py

#!/usr/bin/python3

# Quick and dirty wordle solver
# b is blank square, y is yellow square, g is green square.

from sys import argv
from time import time

# Wordle Solver class
class WordleSolve:

    def __init__(self, file_name, start_word):

        # Alphabet letters sorted by frequency
        self.letter_freq = ["s", "e", "a", "r", "o", "i", "l", "t", "n", "u", "d", "c", "y", "p", "m", "h", "g", "b", "k", "f", "w", "v", "z", "x", "j", "q"]
        self.letter_freq_value = [46.1, 44.8, 40.5, 30.9, 29.5, 28.1, 25, 24, 21.4, 18.6, 18.1, 15.7, 15.4, 14.6, 14.2, 13.3, 11.8, 11.5, 10.2, 7.9, 7.7, 5.2, 2.5, 2.4, 2.1, 0.9]
        self.letter_pairs = ["th", "he", "an", "in", "er", "nd", "re", "ed", "es", "ou", "to", "ha", "en", "ea", "st", "nt", "on", "at", "hi", "as", "it", "ng", "is", "or", "et", "of", "ti", "ar", "te", "se", "me", "sa", "ne", "wa", "ve", "le", "no", "ta", "al", "de", "ot", "so", "dt", "ll", "tt", "el", "ro", "ad", "di", "ew", "ra", "ri", "sh"]
        self.letter_pairs_freq = [33, 30.2, 18.1, 17.9, 16.9, 14.6, 13.3, 12.6, 11.5, 11.5, 11.5, 11.4, 11.1, 11, 10.9, 10.6, 10.6, 10.4, 9.7, 9.5, 9.3, 9.2, 8.6, 8.4, 8.3, 8, 7.6, 7.5, 7.5, 7.4, 6.8, 6.7, 6.6, 6.6, 6.5, 6.4, 6, 5.9, 5.7, 5.7, 5.7, 5.7, 5.6, 5.6, 5.6, 5.5, 5.5, 5.2, 5, 5, 5, 5, 5]
        self.target_word = ["_", "_", "_", "_", "_"]
        self.yellow_green_list = []
        self.not_list = []
        self.word_history = []
        self.score_list = []
        self.start_word = start_word

        # Open a list of 5 letter words and save in memory
        word_list_file = open(file_name, "r")
        self.word_list = word_list_file.read().split()
        word_list_file.close()
        self.word_list_total = len(self.word_list)

    # Print a list of words in a box like Wordle does
    def box(self, words):
        # Box top and bottom
        box_top = "  ___________________ "
        box_spacer  = "  ------------------- "   
        output = box_top
        for word in words:
            output += "\n" + ' | ' +  ' | '.join(list(word)) + ' | ' + "\n" + box_spacer
        return output

    # Get user input
    def get_input(self):
        text = input("Enter results: ")
        input_list = []
        for letter in text:
            input_list.append(letter)
        return input_list

    # Print iterations progress
    def printProgressBar (self, iteration, total, prefix = '', suffix = '', decimals = 1, length = 100, fill = '#', printEnd = "\r"):
        percent = ("{0:." + str(decimals) + "f}").format(100 * (iteration / float(total)))
        filledLength = int(length * iteration // total)
        bar = fill * filledLength + '-' * (length - filledLength)
        print(f'\r{prefix} |{bar}| {percent}% {suffix}', end = printEnd)
        # Print New Line on Complete
        if iteration == total: 
            print()

    # If we know a letter is or isn't in a word, adjust the word list apropriatly
    def letter_modify(self, letter, remove):
        i = 0
        while i < len(self.word_list):
            if remove:
                if letter in self.word_list[i]:
                    if letter not in self.yellow_green_list:
                        del self.word_list[i]
                        del self.score_list[i]
                        continue
            if not remove:
                if letter not in self.word_list[i]:
                    del self.word_list[i]
                    del self.score_list[i]
                    continue
            i += 1

    # If we know what place a letter is or isn't, adjust the word list apropriatly
    def letter_place(self, place, letter, remove):
        i = 0
        while i < len(self.word_list):
            if remove:
                if letter in self.word_list[i][place]:
                    del self.word_list[i]
                    del self.score_list[i]
                    continue
            if not remove:
                if letter not in self.word_list[i][place]:
                    del self.word_list[i]
                    del self.score_list[i]
                    continue
            i += 1

    # Add known letters in the word to a list
    def add_known_letters(self, letter):
        if letter not in self.yellow_green_list:
            self.yellow_green_list.append(letter)

    # Add known letter not in the word to a list
    def add_known_not_letters(self, letter):
        if letter not in self.not_list:
            self.not_list.append(letter)

    # Print a few stats
    def stats(self):
        print("\nWordle: %s" % ' '.join(self.target_word))
        print("Known letters in word: %s" % ' '.join(self.yellow_green_list))
        print("Known letters not in word: %s" % ' '.join(self.not_list))
        print("{0} words remaining, down to {1:.2%} of words.\n".format(len(self.word_list), (len(self.word_list)/self.word_list_total)))

    # Sort word list based on a letter frequency score
    def sort_by_freq(self):
        self.score_list = []
        for word in self.word_list:
            i = 1
            score = 0
            letters = []
            for letter in word:
                # Add to score for top signifigant letters by position
                if i == 1 and letter == "s": score += 11.5
                if i == 2 and letter == "a": score += 12.5
                if i == 3 and letter == "r": score += 11.7
                if i == 4 and letter == "e": score += 12.5
                if i == 5 and letter == "s": score += 13.1
                if letter not in letters:
                    # Add the frequency score of a letter to the score.
                    score += float(self.letter_freq_value[self.letter_freq.index(letter)])
                else:
                    # If a letter is in a word more than once, reduce value of any other occurences. Magic number seems freq/5. 
                    score += float(self.letter_freq_value[self.letter_freq.index(letter)])/5
                letters.append(letter)
                i += 1
            # Now, we add a value based upon frequency of common letter pairs. 
            for pair in self.letter_pairs:
                if pair in word:
                    score += float(self.letter_pairs_freq[self.letter_pairs.index(pair)])
            self.score_list.append(score)
        # Sort the list of scores.
        zipped_list = sorted(zip(self.score_list, self.word_list), reverse=True)
        self.word_list = [self.word_list for score_list, self.word_list in zipped_list]
        self.score_list = sorted(self.score_list, reverse=True)
        return zipped_list

    # Process input and adjust the word list starting with known letters in the target word
    def process_guess(self, input_list, iteration):
        # Known letter and position [Green square]
        i = 0
        for letter in input_list:
            if letter == "g":
                self.add_known_letters(self.word_history[iteration][i])
                self.letter_place(i, self.word_history[iteration][i], False)
                self.target_word[i] = self.word_history[iteration][i]
            i +=1

        # Known letter but not position [Yellow square]
        i = 0
        for letter in input_list:
            if letter == "y":
                self.add_known_letters(self.word_history[iteration][i])
                self.letter_place(i, self.word_history[iteration][i], True)
                self.letter_modify(self.word_history[iteration][i], False)
            i +=1
        
        # Known not a letter in the target word [Blank square]
        i = 0
        for letter in input_list:
            if letter == "b":
                self.letter_modify(self.word_history[iteration][i], True)
                self.add_known_not_letters(self.word_history[iteration][i])
            i += 1

    # Wordle logic for running tests
    def wordle_logic(self, guess, wordle):
        i = 0
        output = []
        for letter in guess:
            if letter == wordle[i]:
                output.append("g")
            elif letter in wordle:
                output.append("y")
            elif letter not in wordle:
                output.append("b")
            else:
                print("Something bad happened.")
            i += 1
        if len(output) != 5:
            print("Something bad happened x2.")
        return output

# Automatically play Wordle for testing code, start words, etc.
def play_bot(solver, target_list, start_word, verbose):
    succsess = []
    solver.word_list = []
    word_list_sorted = []
    score_list_sorted = []
    fails = 0
    for word in target_list:
        solver.word_list.append(word)

    # Sort by frequency and backup those values
    scores = solver.sort_by_freq()
    for word in solver.word_list:
        word_list_sorted.append(word)
    for score in solver.score_list:
        score_list_sorted.append(score)
    j = 0
    for wordle in target_list:
        if verbose: solver.printProgressBar(j, len(target_list) - 1, prefix = start_word, suffix = "Done", length = 50)
        j += 1
        solver.word_list = []
        solver.word_history = []
        solver.yellow_green_list = []
        solver.score_list = [] 

        # Copy the saved frequency values instead of calculating them again
        for word in word_list_sorted:
            solver.word_list.append(word)
        for score in score_list_sorted:
            solver.score_list.append(score)

        i = 0
        for i in range(0, 6):
            if i == 0:
                solver.word_history.append(start_word)
                guess = solver.wordle_logic(start_word, wordle)
            else:
                solver.word_history.append(solver.word_list[0])
                guess = solver.wordle_logic(solver.word_list[0], wordle)
            solver.process_guess(guess, i)
            if ''.join(guess) == "ggggg":
                succsess.append(i + 1)
                break
            if len(solver.word_list) == 1:
                succsess.append(i + 1)
                break
            if i == 5:
                fails += 1
            if solver.word_list[0] in solver.word_history:
                solver.word_list.pop(0)
    count = 0
    for i in succsess:
        count += i
    return count, succsess, fails


# Using each word in the word list as a Wordle, try to solve it and see what the average number of guesses is
def test():
    start_word = "least"
    successes = []
    fails = 0
    target_list = []
    solver = WordleSolve("5_letter_word_list.txt", start_word)
    for word in solver.word_list:
        target_list.append(word)
    count, successes, fails = play_bot(solver, target_list, start_word, True)
    print("Number of wordles: %s" % len(target_list))
    print("Number of succsess: %s" % len(successes))
    print("Average guesses for successes: %s" % (count/len(successes)))
    print("Fails: %s" % fails)

# Using each word in the word list as a Wordle, try to solve it and see what the average number of guesses is
def test_start_words():
    averages = []
    target_list = []
    start_word_list =[]
    start_word = "least"
    solver = WordleSolve("5_letter_word_list.txt", start_word)
    for word in solver.word_list:
        start_word_list.append(word)
    for word in solver.word_list:
        target_list.append(word)
    for start_word in start_word_list:
        successes = []
        fails = 0
        count, successes, fails = play_bot(solver, target_list, start_word, True)
        averages.append((count/len(successes), start_word))
    result = sorted(averages)
    print("The best start word is %s with a average of %f Wordle guesses." % (result[0][1], result[0][0]))

# Solve a Wordle with a user
def play():
    start_word = "least"
    solver = WordleSolve("5_letter_word_list.txt", start_word)
    solver.stats()
    print("Begin with:")
    scores = solver.sort_by_freq()
    print(solver.box([start_word]))
    solver.word_history.append(start_word)
    i = 0
    for i in range(0, 6):
        user_input = solver.get_input()
        solver.process_guess(user_input[:5], i)
        solver.stats()
        print("Top words:")
        try:
            for x in range(0, 4):
                print("""%s (Freq score: %.2f)""" % (solver.word_list[x], round(solver.score_list[x], 2)))
        except Exception as e:
            pass
        if len(solver.word_list) == 1:
            if ''.join(user_input[:5]) == "ggggg":
                print(solver.box(solver.word_history))
            else:
                print(solver.box(solver.word_history + [solver.word_list[0]]))
            print("Done! Hooray!!")
            exit()
        if solver.word_list[0] in solver.word_history:
            solver.word_list.pop(0)
        print(solver.box(solver.word_history + [solver.word_list[0]]))
        solver.word_history.append(solver.word_list[0])
    print("Oof. Good luck picking final word!")

# Main
if __name__ == "__main__":
    startTime = time()
    if len(argv) == 1:
        print("Try again.")
    elif argv[1] == "test":
        test()
    elif argv[1] == "play":
        play()
    elif argv[1] == "test_start_words":
        test_start_words()
    else:
        print("Try again.")
    executionTime = (time() - startTime)
    print('Time: ' + str(executionTime))