Serbilian cyrilic: a cyrilic script for Brazilian Portuguese based in Serbian Cyrilic

serbilian.rb

#!/usr/bin/ruby
require 'logger'
require 'pragmatic_tokenizer'
require 'unicode_utils'

class Serbilian
  SINGLE_LETTER_WORDS = {
    "à" => "a", 
    "ã" => "ɐ̃", 
    "â" => "ɐ̃", 
    "é" => "э", 
    "ê" => "e",
    "o" => "у",
    "ó" => "ɔ", 
    "ô" => "o", 
    "õ" => "õ", 
    "í" => "i", 
    "u" => "у"
  } 
  PHONEME_TABLE = {
    ""   => ["", ""],
    "a"  => ["А", "а"],
    "æ"  => ["А", "а"],
    "ɐ"  => ["∀", "ɐ"],
    "ɐ̃"  => ["∀", "ɐ"],
    "ã"  => ["∀", "ɐ"],
    "aɪ" => ["Aи", "aи"],
    "ɐ̃ʊ̃" => ["∀o", "ɐo"],
    "aʊ" => ["Aу", "aу"],
    "b"  => ["Б", "б"],
    "d"  => ["Д", "д"],
    "dʒ" => ["Џ", "џ"],
    "e"  => ["Е", "е"],
    "eɪ" => ["E", "e"],
    "ɛ"  => ["Э", "э"],
    "ə"  => ["", ""],
    "f"  => ["Ф", "ф"],
    "ɡ"  => ["Г", "г"],
    "i"  => ["И", "и"],
    "iʊ" => ["Иу", "иу"],
    "ɪ"  => ["Ј", "ј"],
    "j"  => ["Ј", "ј"],
    "ʒ"  => ["Ж", "ж"],
    "k"  => ["К", "к"],
    "l"  => ["Л", "л"],
    "ʎ"  => ["Љ", "љ"],
    "m"  => ["М", "м"],
    "n"  => ["Н", "н"],
    "ŋ"  => ["Н", "н"],
    "ɲ"  => ["Њ", "њ"],
    "o"  => ["О", "о"],
    "õ"  => ["Õ", "õ"],
    "oɪ" => ["Oи", "oи"],
    "ɔ"  => ["Ɔ", "ɔ"],
    "p"  => ["П", "п"],
    "ɾ"  => ["Р", "р"],
    "r"  => ["Р", "р"],
    "x"  => ["X", "x"],
    "ʁ"  => ["X", "x"],
    "s"  => ["С", "с"],
    "ʃ"  => ["Ш", "ш"],
    "t"  => ["Т", "т"],
    "tʃ" => ["Ч", "ч"],
    "ũ"  => ["У", "у"],
    "ũ"  => ["У", "у"],
    "ʊ"  => ["У", "у"],
    "ʊ̃"  => ["У", "у"],
    "у"  => ["У", "у"],
    "v"  => ["В", "в"],
    "w"  => ["у", "у"],
    "y"  => ["I", "i"],
    "z"  => ["S", "s"],
  }

  TOKENIZER = PragmaticTokenizer::Tokenizer.new :language => "pt-br", :downcase => false

  LOGGER = Logger.new(STDOUT)
  LOGGER.level = Logger::INFO

  def self.word?(token)
    not token.match(/[[:word:]]+/).nil?
  end

  def self.downcase(letter)
    UnicodeUtils.downcase(letter, :pt)
  end

  def self.word_to_ipa(word)
    single_letter_replacement = SINGLE_LETTER_WORDS[downcase(word.strip)]
    if single_letter_replacement
      single_letter_replacement
    else
      %x{espeak -v pt-br --ipa=3 "#{word}" -q}.strip
    end
  end

  def self.ipa_to_serbilian(ipa, capitalize)
    serbilian = ""
    ipa.split("_").each_with_index do |s, i| 
      clean_phoneme = s.tr("ˈ", "").tr("ˌ", "")
      chars = PHONEME_TABLE[clean_phoneme]
      LOGGER.warn ("missing phoneme table entry for /#{clean_phoneme}/") unless chars
      lcase = (capitalize and i.zero?) ? 0 : 1
      serbilian << chars[lcase] if chars
    end
    serbilian
  end

  def self.upcase?(letter)
    UnicodeUtils.upcase(letter, :pt) == letter
  end

  def self.word_to_serbilian(word)
    ipa = word_to_ipa(word)
    serbilian = ipa_to_serbilian(ipa, upcase?(word[0]))
    LOGGER.debug("'#{word}' -> '#{ipa}' -> '#{serbilian}'")
    serbilian 
  end

  def self.join_tokens(tokens)
    res = ""
    tokens.each_with_index do |token, i|
      res << token
      next_token = tokens[i + 1] if i + 1 < tokens.length
      res << " " if next_token and word?(next_token) 
    end
    res
  end

  def self.ptbr_to_serbilian(ptbr)
    LOGGER.debug("pt-br: #{ptbr}")
    tokens = TOKENIZER.tokenize(ptbr).map do |token|
      word?(token) ? word_to_serbilian(token) : token
    end
    join_tokens(tokens)
  end

  def self.test_pangrams
    pangrams = [
      "Um pequeno jabuti xereta viu dez cegonhas felizes.",
      "Blitz prende ex-vesgo com cheque fajuto.",
      "Gazeta publica hoje no jornal uma breve nota de faxina na quermesse.",
      "Zebras caolhas de Java querem passar fax para moças gigantes de New York.",
      "Luís argüia à Júlia que brações, fé, chá, óxido, pôr, zângão eram palavras do português.",
      "À noite, vovô Kowalsky vê o ímã cair no pé do pingüim queixoso e vovó põe açúcar no chá de tâmaras do jabuti feliz."
    ]

    pangrams.map { |p| ptbr_to_serbilian(p) }
  end
end