md5_database_verify

#!/usr/bin/env python
# -*- coding: utf-8 -*-
# Filename: md5_database_verify.py
# Version: 1.0.2
# Author: Jeoi Reqi

"""
Description:
    - This script allows users to verify MD5 sums against a database.
    - It loads an MD5 dictionary from a file, prompts the user to input a string & computes the MD5 sum of the input string.
    - It then checks if the user input string already exists in the dictionary.
    - If the string is not found, the user has the option to save it to the database.

Functions:
    - md5sum(string):
        Returns the MD5 sum of a given string.
    - load_md5_database(filename):
        Loads the MD5 dictionary from a file.
    - save_md5_database(md5_dict, filename):
        Saves the MD5 dictionary to a file.
    - main():
        Main function to execute the script.

Requirements:
    - Python 3.x

Usage:
    - Run the script and follow the prompts. Enter a string to look up its MD5 sum, or type 'exit' to quit.
    - If the string is not found in the database, you will be prompted to save it.

Additional Notes:
    - Make sure the database file exists and is accessible.
    - The database file will be overwritten if any changes are made.
"""

import hashlib
import json
import os

def md5sum(string):
    """
    Return the MD5 sum of a given string.

    Args:
        string (str): The input string for which the MD5 sum will be calculated.

    Returns:
        str: The MD5 sum of the input string.
    """
    return hashlib.md5(string.encode()).hexdigest()

def load_md5_database(filename):
    """
    Load the MD5 dictionary from a file.

    Args:
        filename (str): The filename of the file containing the MD5 dictionary.

    Returns:
        dict: The loaded MD5 dictionary.
    """
    md5_dict = {}
    with open(filename, 'r', encoding='utf-8') as infile:
        for line in infile:
            # Extract string and MD5 hash from each line
            parts = line.strip().split(':', 1)
            if len(parts) == 2:
                string, md5_hash = parts
                # Remove quotes and whitespace from string
                string = string.strip().strip('"')
                # Remove quotes, whitespace, and trailing comma from MD5 hash
                md5_hash = md5_hash.strip().strip('"').strip(',')
                # Add to the dictionary
                md5_dict[string] = md5_hash
    return md5_dict

def save_md5_database(md5_dict, filename):
    """
    Save the MD5 dictionary to a file.

    Args:
        md5_dict (dict): The MD5 dictionary to be saved.
        filename (str): The filename of the file where the MD5 dictionary will be saved.
    """
    with open(filename, 'w', encoding='utf-8') as outfile:
        for string, md5_hash in sorted(md5_dict.items()):
            outfile.write(f'"{string}": "{md5_hash}",\n')

def main():
    # Load the MD5 dictionary
    database_filename = 'sum_md5_dict.txt'
    md5_dict = load_md5_database(database_filename)

    while True:
        # Get user input
        user_input = input("\nEnter a string to look up its MD5 sum (or type 'exit' to quit): ").strip()

        # Check if the user wants to exit
        if user_input.lower() == 'exit':
            print("\nExiting the program...\tGoodBye!\n")
            break

        # Compute the MD5 sum of the user input
        user_md5 = md5sum(user_input)

        print("\nComputed MD5 sum:")
        print(user_md5)

        # Check if the MD5 sum exists in the dictionary
        if user_input in md5_dict:
            print(f"\nThe user string '{user_input}' was found in the database.")
            print(f"\nUser string: {user_input}")
            print(f"MD5 sum: {md5_dict[user_input]}")
        else:
            print(f"\nThe user string '{user_input}' was not found in the database.")
            # Ask the user if they want to save the string and its MD5 sum to the database
            save_option = input("\nDo you want to save this string and its MD5 sum to the database?\n1: Yes\n2: No\n\nMake your selection (1 or 2): ").strip()
            if save_option == '1':
                md5_dict[user_input] = user_md5
                save_md5_database(md5_dict, database_filename)
                print("\nString and its MD5 sum saved to the database.")
            else:
                print("\nString and its MD5 sum not saved to the database!")

if __name__ == "__main__":
    main()