'''
    DS2000
    Spring 2020
    Source code - processing a messy file
'''

PUNCTUATION = '\'\",.?!-_;'

def process_file(filename):
    ''' Function: process_file
        Parameter: string
        Returns: contents of the file
    '''
    with open(filename, 'r') as infile:
        contents = infile.readlines()
    return contents

def clean_file(contents):
    ''' Function: clean_file
        Parameters: list of strings
        Return: list of strings, but better
    '''
    clean_list = []
    for line in contents:
        clean = line.strip()
        clean = clean.lower()
        for p in PUNCTUATION:
            clean = clean.replace(p, ' ')
        clean_list.append(clean)
        
    return clean_list
    

def main():

    # Step One: Read the file as-is
    file_contents = process_file('khoury.txt')

    # Step Two: Clean the file by removing whitespace and punctuation
    cleaner = clean_file(file_contents)

    # Step Three: Turn the list of sentences into a list of words
    words = ' '.join(cleaner)
    words = words.split()
    print(words)

main()