Readable Caesar cipher

I was trying to refactor the following code for Caesar cipher, but it feels overly complicated at the end of the refactoring. Could someone else take a look at it and let me know if I am just overdoing/making it complicated for someone trying to read the code?

My highest priority is to write readable code, so any pointers in that regard would be welcome.

module CASE_WRAP
  def cyclic_next!
    if self == 'z'
      replace 'a'
      return
    end

    if self == 'Z'
      replace 'A'
      return
    end

    self.next!
  end

  def shift(number)
    number.times { self.cyclic_next! }
    self
  end
end

class CaesarCipher
  def self.cipher(message, shift)
    CaesarCipher.new(shift: shift)
      .encrypt message
  end

  def initialize(shift: 0)
    @shift = shift
  end

  def encrypt(message)
    message.gsub(/\w/) do |character|
      character.extend(CASE_WRAP)
        .shift @shift
    end
  end
end

Find tests for this in my repo.

All the caesar cypher does is rotate an alphabet by n places. If we let the alphabet be an Array, we can use Array#rotate to do the rotating for us. To substitute single characters, there is String#tr. Its name and usage are a little odd, but you'll quickly get used to it.

Basically, it maps characters in a String to characters in another String by their index. So "abc".tr("abc", "123") maps "a" to "1", "b" to "2" and "c" to "3".It expands ranges such as "a-c" first, so you can can write "abc".tr("abc", "123") as "abc".tr("a-c","1-3").

Though the interface is different, here's a simple implementation of the Caesar cypher.

module Caesar
  def self.encrypt(message, shift = 3)
    message.tr("a-zA-Z", rotated_alphabet(shift))
  end

  def self.decrypt(message, shift = 3)
    encrypt(message, 26 - shift)
  end

  def self.rotated_alphabet(shift)
    lower = "a".upto("z").to_a.rotate(shift)
    upper = "A".upto("Z").to_a.rotate(shift)

    lower.concat(upper).join
  end
end

Caesar.decrypt Caesar.encrypt("hi")
# => "hi"

For efficiency, you could change how you calculate the rotated alphabets. Here's an efficient, but not exactly a readable implementation.

def self.rotated_alphabet(shift)
    lower = (97 + shift).chr # 97 is the ASCII code for "a"
    upper = (65 + shift).chr # 65 is the ASCII code for "A"

    "#{lower}-za-#{lower}#{upper}-ZA-#{upper}"
  end

StackExchange Code Review Q#80123, answer score: 2