preprocessing.py

"""
preprocessing.py
This module handles reading the ciphertext and cleaning it up
"""


def load_ciphertext(filepath):
    """
    Reads the ciphertext file and returns it as a string
    """
    with open(filepath, 'r', encoding='utf-8') as f:
        return f.read()


def preprocess_text(raw_text):
    """
    Removes all non-alphabetic characters and converts to uppercase
    Also records where spaces and punctuation were so we can put them back later
    Returns: (clean_text, non_alpha_positions)
    """
    clean_text = ""
    non_alpha_positions = []
    
    # Go through each character
    for char in raw_text:
        if char.isalpha():
            # Convert to uppercase and treat J as I (Playfair rule)
            letter = char.upper()
            if letter == 'J':
                letter = 'I'
            clean_text += letter
        else:
            # Record where this punctuation/space was
            # Position is where it should go back in the clean text
            position = len(clean_text)
            non_alpha_positions.append((position, char))
    
    return clean_text, non_alpha_positions


def restore_formatting(decrypted_text, non_alpha_positions):
    """
    Puts back all the spaces and punctuation that we removed earlier
    """
    # Start with the decrypted text
    result = decrypted_text
    
    # Insert punctuation back in reverse order so positions don't shift
    # Sort by position in reverse
    sorted_positions = sorted(non_alpha_positions, key=lambda x: x[0], reverse=True)
    
    for position, char in sorted_positions:
        # Insert the character at the recorded position
        result = result[:position] + char + result[position:]
    
    return result