Vigenère cipher in Java for all UTF-8 characters

Question

I have this simple function for encrypting strings via Vigenère in Java. I omitted the decryption as this is just a "-" instead of the "+" in the line where the new value is calculated.

But this function works only for the normal alphabet A-Z. How can I change the function so that it supports lowercase letters as well as uppercase letters and all other UTF-8 chars?

public static String vigenere_encrypt(String plaintext, String key) {
    String encryptedText = "";
    for (int i = 0, j = 0; i < plaintext.length(); i++, j++) {
        if (j == key.length()) { j = 0; } // use key again if end reached
        encryptedText += (char) ((plaintext.charAt(i)+key.charAt(j)-130)%26 + 65);
    }
    return encryptedText;
}

Thank you very much for your help!

Answer 1

Another answer, that does do the Vigenere cipher on upper & lower case characters, simply inserting the other characters. Use this technique to create multiple groups of characters to encode.

public static String vigenere(String plaintext, String key, boolean encrypt) {

    final int textSize = plaintext.length();
    final int keySize = key.length();

    final int groupSize1 = 'Z' - 'A' + 1; 
    final int groupSize2 = 'z' - 'a' + 1;
    final int totalGroupSize = groupSize1 + groupSize2;

    final StringBuilder encryptedText = new StringBuilder(textSize);
    for (int i = 0; i < textSize; i++) {
        final char plainChar = plaintext.charAt(i);

        // this should be a method, called for both the plain text as well as the key
        final int plainGroupNumber; 
        if (plainChar >= 'A' && plainChar <= 'Z') {
            plainGroupNumber = plainChar - 'A';
        } else if (plainChar >= 'a' && plainChar <= 'z') {
            plainGroupNumber = groupSize1 + plainChar - 'a';
        } else {
            // simply leave spaces and other characters
            encryptedText.append(plainChar);
            continue;
        }

        final char keyChar = key.charAt(i % keySize);
        final int keyGroupNumber; 
        if (keyChar >= 'A' && keyChar <= 'Z') {
            keyGroupNumber = keyChar - 'A';
        } else if (keyChar >= 'a' && keyChar <= 'z') {
            keyGroupNumber = groupSize1 + keyChar - 'a';
        } else {
            throw new IllegalStateException("Invalid character in key");
        }

        // this should be a separate method
        final int cipherGroupNumber;
        if (encrypt) {
            cipherGroupNumber = (plainGroupNumber + keyGroupNumber) % totalGroupSize;
        } else {
            // some code to go around the awkward way of handling % in Java for negative numbers
            final int someCipherGroupNumber = plainGroupNumber - keyGroupNumber;
            if (someCipherGroupNumber < 0) {
                cipherGroupNumber = (someCipherGroupNumber + totalGroupSize);
            } else {
                cipherGroupNumber = someCipherGroupNumber;
            }
        }

        // this should be a separate method
        final char cipherChar;
        if (cipherGroupNumber < groupSize1) {
            cipherChar = (char) ('A' + cipherGroupNumber);
        } else {
            cipherChar = (char) ('a' + cipherGroupNumber - groupSize1);
        }
        encryptedText.append(cipherChar);
    }

    return encryptedText.toString();
}

Again, this is unsafe code as the cipher used has been broken for ages. Don't use too many 'A' characters in your keys :) But the character encoding should be sound.

Answer 2

Well, you asked for it and I felt like puzzling, but print out the cipher text and you will know what you just asked for...

public static String vigenereUNICODE(String plaintext, String key, boolean encrypt) {

    final int textSize = plaintext.length();
    final int keySize = key.length();

    final StringBuilder encryptedText = new StringBuilder(textSize);
    for (int i = 0; i < textSize; i++) {
        final int plainNR = plaintext.codePointAt(i);
        final int keyNR = key.codePointAt(i % keySize);

        final long cipherNR;
        if (encrypt) {
            cipherNR = ((long) plainNR + (long) keyNR) & 0xFFFFFFFFL;
        } else {
            cipherNR = ((long) plainNR - (long) keyNR) & 0xFFFFFFFFL;
        }

        encryptedText.appendCodePoint((int) cipherNR);
    }

    return encryptedText.toString();
}

EDIT: Please don't ever use in production code, as I haven't got a clue if the code points can indeed be encoded/decoded. Not all points have been defined, as far as I know, and the standard is a moving target.

Answer 3

If full unicode support is not possible and you have to define your list of valid characters, anyway, why not just use a function like this?

public static String vigenere_cipher(String plaintext, String key, boolean encrypt) {

    String alphabet = "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ ,.-"; // including some special chars
    final int alphabetSize = alphabet.length();
    final int textSize = plaintext.length();
    final int keySize = key.length();
    final StringBuilder encryptedText = new StringBuilder(textSize);

    for (int i = 0; i < textSize; i++) {
        final char plainChar = plaintext.charAt(i); // get the current character to be shifted
        final char keyChar = key.charAt(i % keySize); // use key again if the end is reached
        final int plainPos = alphabet.indexOf(plainChar); // plain character's position in alphabet string
        if (plainPos == -1) { // if character not in alphabet just append unshifted one to the result text
            encryptedText.append(plainChar);
        }
        else { // if character is in alphabet shift it and append the new character to the result text
            final int keyPos = alphabet.indexOf(keyChar); // key character's position in alphabet string
            if (encrypt) { // encrypt the input text
                encryptedText.append(alphabet.charAt((plainPos+keyPos) % alphabetSize));
            }
            else { // decrypt the input text
                int shiftedPos = plainPos-keyPos;
                if (shiftedPos < 0) { // negative numbers cannot be handled with modulo
                    shiftedPos += alphabetSize;
                }
                encryptedText.append(alphabet.charAt(shiftedPos));
            }
        }
    }

    return encryptedText.toString();

}

This should be a very short and working version. And the alphabet can easily be stored in a string that can always be extended (which results in different ciphertexts).