A Ruby helper method to convert CJK (Japanese) Zenkaku to ASCII

zenkaku_to_ascii.rb

# coding: utf-8

# Nkf is a built-in Gem/library up to Ruby 3.0 at least.
# 
# The following method is a wrapper for NKF.
# 
# NKF (Nihon-go Kanji Filter) is an old library since 1990s. But I find it is
# the most convenient library still in 2021 to handle text with
# some nasty Japanese character code (n.b., the "proper" Japanese character code
# sets are fine, but unfortunately the character code sets popularalized
# by Microsoft and Apple though its base had been born and used before
# are pretty irregular...).
require 'nkf'

  # Method to convert (CJK) Zenkaku alphabet/number/symbol to Hankaku.
  #
  # A JIS space is converted to 2 ASCII spaces in default (option zopt == 2),
  # corresponding to the NKF option '-Z'.
  # The other NKF options should be given as a string: opts.
  # Option '-w' is in default, which means the output is in UTF-8,
  # unless one of [-j, -e, -s] (JIS, EUC, SJIS) is included in opts.
  # Also opton '-m0' is default.
  #
  # @example
  #   zenkaku_to_ascii('（あ）', zopt: 1)  # => '(あ)'
  #
  # @param instr [String] Input to NKF
  # @option opts [String] NKF option string
  # @param zopt: [Integer, NilClass]
  # @return [String]
  def zenkaku_to_ascii(instr, opts='', zopt: nil)
    z_spaces = (zopt || 2)

    if /(^| )-[jesw]/ !~ opts
      optsnkf = ("-w "+opts).strip
    end
    NKF.nkf("-m0 -Z#{z_spaces} #{optsnkf}", instr)   # [-Z2] Convert a JIS X0208 space to 2 ASCII spaces, as well as Zenkaku alphabet/number/symbol to Hankaku.
  end

############ Tests ############

if $0 == __FILE__
  require 'minitest/autorun'

  class TestZenkakuToAsciiClass < MiniTest::Test
    def setup; end
    def teardown; end
    def test_zenkaku_to_ascii
      assert_equal '(あ)', zenkaku_to_ascii('（あ）', zopt: 1)
      assert_equal '1 A',  zenkaku_to_ascii("\uff11\u3000\uff21", zopt: 1)
      assert_equal '1  A', zenkaku_to_ascii("\uff11\u3000\uff21", zopt: 2)
    end
  end
end