Methods to split a Ruby string into continuous CJK (Chinese/Japanese/Korean) and alphabetical string segments.

string_extensions.rb

# encoding: UTF-8

class String

  def contains_cjk?
    !!(self =~ /\p{Han}|\p{Katakana}|\p{Hiragana}|\p{Hangul}/)
  end

  def for_each_continous_cjk_alpha_part
    split_into_cjk_alpha_phrases.each.with_index do |part, i|
      yield part, i if block_given?
    end
  end

  def split_into_cjk_alpha_phrases
    identify_cjk_alpha_ranges.map {|range|
      self[range]
    }
  end

  private
    def identify_cjk_alpha_ranges
      split_index = identify_cjk_alpha_split_indexes
      capped_split_index = case
        when split_index.first !=0 && split_index.last != (length - 1)
          [0] + split_index + [(length - 1)]
        when split_index.first !=0
          [0] + split_index
        when split_index.last != (length - 1)
          split_index + [(length - 1)]
        else
          split_index
        end
      capped_split_index.each_cons(2).map {|(x, y)|
        if x == 0
          (x .. y)
        else
          (x+1 .. y)
        end
      }
    end

    def identify_cjk_alpha_split_indexes
      consecutive_pairs = split('').each_cons(2).to_a
      consecutive_pairs.map.with_index{|(x, y), i|
        if x.contains_cjk? != y.contains_cjk?
          i
        end
      }.reject(&:nil?)
    end
end