gimchi 0.1.0 → 0.1.1

data/README.ko.rdoc

@@ -1,5 +1,3 @@
-# encoding: UTF-8
-
 = gimchi
 
 == 개요
@@ -13,9 +11,14 @@ Gimchi는 한글 스트링을 다롭니다.
 - 한글을 초성, 중성, 종성으로 분리하고, 이를 다시 합치는 기능
 - 숫자 표기를 한글 표현으로 변환
 
+== 설치
+  gem install gimchi
+
 == 사용법
 
 === Gimchi::Korean 인스턴스의 생성
+  require 'gimchi'
+  
   ko = Gimchi::Korean.new
 
 === 한글 캐릭터 여부 판단
@@ -54,7 +57,6 @@ Gimchi는 한글 스트링을 다롭니다.
 === 숫자 읽기
   ko.read_number(1999)         # "천 구백 구십 구"
   ko.read_number(- 100.123)    # "마이너스 백점일이삼"
-  ko.read_number("153,101,202,333.321")
   ko.read_number("153,191,100,678.3214")
     	# "천 오백 삼십 일억 구천 백 십만 육백 칠십 팔점삼이일사"
 
@@ -87,7 +89,7 @@ Gimchi는 한글 스트링을 다롭니다.
   ko.romanize str, :as_pronounced => false
     # "Dwaet-eo dwaet-eo ije geureon gareuchim-eun dwaet-eo mae-il achim ilgop si samsip bunkkaji uril jogeuman gyosillo mol-aneogo"
   ko.romanize str, :number => false
-	# "Dwaesseo dwaesseo ije geureon gareuchimeun dwaesseo mae-il achim 7 si 30 bunkkaji uril jogeuman gyosillo moraneoko"
+    # "Dwaesseo dwaesseo ije geureon gareuchimeun dwaesseo mae-il achim 7 si 30 bunkkaji uril jogeuman gyosillo moraneoko"
 
 == 구현의 한계
 

data/README.rdoc

@@ -1,5 +1,3 @@
-# encoding: UTF-8
-
 = gimchi
 
 Gimchi is a simple Ruby gem which knows how to handle Korean strings. It knows
@@ -10,11 +8,16 @@ and how they're written in roman alphabet.
 Gimchi (only partially) implements the following rules dictated by
 The National Institute of The Korean Language (http://www.korean.go.kr)
 * Korean Standard Pronunciation
-* Korean romanization
+* Korean Romanization
+
+== Installation
+  gem install gimchi
 
 == Usage
 
 === Creating Gimchi::Korean instance
+  require 'gimchi'
+  
   ko = Gimchi::Korean.new
 
 === Checks if the given character is in Korean alphabet
@@ -53,7 +56,6 @@ The National Institute of The Korean Language (http://www.korean.go.kr)
 === Reading numbers in Korean
   ko.read_number(1999)         # "천 구백 구십 구"
   ko.read_number(- 100.123)    # "마이너스 백점일이삼"
-  ko.read_number("153,101,202,333.321")
   ko.read_number("153,191,100,678.3214")
     	# "천 오백 삼십 일억 구천 백 십만 육백 칠십 팔점삼이일사"
 
@@ -91,13 +93,13 @@ The National Institute of The Korean Language (http://www.korean.go.kr)
 == Limitation of the implementation
 
 Unfortunately in order to implement the complete specification of Korean
-pronunciation and romanization, we need NLP, hugh Korean dictionaries and even
+pronunciation and romanization, we need NLP, huge Korean dictionaries and even
 semantic analysis of the given string. And even with all those complex
 processing, we cannot guarantee 100% accuracy of the output. So yes, that is
 definitely not what this gem tries to achieve. Gimchi tries to achieve "some"
 level of accuracy with relatively simple code.
 
-Currently, Gimchi code containts a lot of ad-hoc (possibly invalid) patches
+Currently, Gimchi code contains a lot of ad-hoc (possibly invalid) patches
 that try to improve the quality of the output, which should better be
 refactored anytime soon.
 

data/config/default.yml

@@ -56,7 +56,11 @@ pronouncer:
         ㅎ:
     transformation:
         # changing the order affects the quality of the transformation
-        sequence:
+        sequence for 1:
+            - rule_5_1
+            - rule_5_3
+
+        sequence for 2:
             - rule_16
             - rule_17
             - rule_18
@@ -96,7 +100,6 @@ number:
         digits: ["", 한, 두, 세, 네, 다섯, 여섯, 일곱, 여덟, 아홉]
         post substitution:
             물살: 무살
-            물시: 무살
 
 romanization:
     chosung:
@@ -143,7 +146,7 @@ romanization:
         ㅢ: ui
     jongsung:
         ㄱ: k
-        ㄴ: n
+        ㄴ: n-
         ㄷ: t
         ㄹ: l
         ㅁ: m
@@ -153,4 +156,14 @@ romanization:
         # 제2항 [붙임 2]‘ㄹ’은 모음 앞에서는 ‘ｒ’로, 자음 앞이나 어말에서는
         # ‘ｌ’로 적는다. 단, ‘ㄹㄹ’은 ‘ll’로 적는다.
         lr: ll
-
+        kkk: k-kk
+        ttt: t-tt
+        ppp: p-pp
+        "--": "-"
+        ? !ruby/regexp /n-([^gaeiou])/
+        : "n\\1"
+        ? !ruby/regexp /-(\s)/
+        : "\\1"
+        ? !ruby/regexp /-$/
+        : ""
+        

data/lib/gimchi/char.rb

@@ -2,15 +2,20 @@
 
 module Gimchi
 class Korean
-	# Class representing each Korean character.
-	# chosung, jungsung and jongsung can be get and set.
+	# Class representing each Korean character. Its three components, 
+	# `chosung', `jungsung' and `jongsung' can be get and set.
 	#
-	# to_s merges components into a String.
-	# to_a returns the three components.
+	# `to_s' merges components into a String. `to_a' returns the three components.
 	class Char
-		attr_reader :org
-		attr_reader :chosung, :jungsung, :jongsung
+		# @return [String] Chosung component of this character.
+		attr_reader :chosung
+		# @return [String] Jungsung component of this character.
+		attr_reader :jungsung
+		# @return [String] Jongsung component of this character.
+		attr_reader :jongsung
 
+		# @param [Gimchi::Korean] kor Gimchi::Korean instance
+		# @param [String] kchar Korean character string
 		def initialize kor, kchar
 			raise ArgumentError('Not a korean character') unless kor.korean_char? kchar
 
@@ -28,16 +33,17 @@ class Korean
 				self.chosung = @kor.chosungs[n1]
 				self.jungsung = @kor.jungsungs[n2]
 				self.jongsung = ([nil] + @kor.jongsungs)[n3]
-			elsif (@kor.chosungs + @kor.jongsungs).include? kchar
+			elsif @kor.chosungs.include? kchar
 				self.chosung = kchar
 			elsif @kor.jungsungs.include? kchar
 				self.jungsung = kchar
+			elsif @kor.jongsungs.include? kchar
+				self.jongsung = kchar
 			end
-
-			@org = self.dup
 		end
 
-		# recombine components into korean character
+		# Recombines components into a korean character.
+		# @return [String] Combined korean character
 		def to_s
 			if chosung.nil? && jungsung.nil?
 				""
@@ -52,49 +58,61 @@ class Korean
 			end
 		end
 
+		# Sets the chosung component.
+		# @param [String] 
 		def chosung= c
 			raise ArgumentError.new('Invalid chosung component') if
 					c && @kor.chosungs.include?(c) == false
 			@chosung = c && c.dup.extend(Component).tap { |e| e.kor = @kor }
 		end
 
+		# Sets the jungsung component
+		# @param [String] 
 		def jungsung= c
 			raise ArgumentError.new('Invalid jungsung component') if
 					c && @kor.jungsungs.include?(c) == false
 			@jungsung = c && c.dup.extend(Component).tap { |e| e.kor = @kor }
 		end
 
+		# Sets the jongsung component
+		#
+		# @param [String] 
 		def jongsung= c
 			raise ArgumentError.new('Invalid jongsung component') if
 					c && @kor.jongsungs.include?(c) == false
 			@jongsung = c && c.dup.extend(Component).tap { |e| e.kor = @kor }
 		end
 
-		# to_a returns the three components.
+		# Returns Array of three components.
+		#
+		# @return [Array] Array of three components
 		def to_a
 			[chosung, jungsung, jongsung]
 		end
 
-		# Check if this is a complete Korean character
+		# Checks if this is a complete Korean character.
 		def complete?
 			chosung.nil? == false && jungsung.nil? == false
 		end
 
-		# Check if this is a non-complete Korean character
+		# Checks if this is a non-complete Korean character.
 		# e.g. ㅇ, ㅏ
 		def partial?
 			chosung.nil? || jungsung.nil?
 		end
 
 	private
-		# nodoc #
+		# Three components of Korean::Char are extended to support #vowel? and #consonant? method.
 		module Component
+			# @return [Korean] Hosting Korean instance
 			attr_accessor :kor
 
+			# Is this component a vowel?
 			def vowel?
 				kor.jungsungs.include? self
 			end
 
+			# Is this component a consonant?
 			def consonant?
 				self != 'ㅇ' && kor.chosungs.include?(self)
 			end
@@ -102,3 +120,4 @@ class Korean
 	end#Char
 end#Korean
 end#Gimchi
+

data/lib/gimchi/korean.rb

@@ -5,35 +5,41 @@ class Korean
 	DEFAULT_CONFIG_FILE_PATH = 
 		File.dirname(__FILE__) + '/../../config/default.yml'
 
+	# Returns the YAML configuration used by this Korean instance.
+	# @return [String]
 	attr_reader   :config
-	attr_accessor :pronouncer
 
 	# Initialize Gimchi::Korean.
-	# You can override many part of the implementation with customized config file.
+	# @param [String] config_file You can override many parts of the implementation by customizing config file
 	def initialize config_file = DEFAULT_CONFIG_FILE_PATH
 		require 'yaml'
 		@config = YAML.load(File.read config_file)
 		@config.freeze
 
-		@pronouncer = Korean::Pronouncer.new(self)
+		@pronouncer = Korean::Pronouncer.send :new, self
 	end
 
-	# Array of chosung's
+	# Array of chosung's.
+	#
+	# @return [Array] Array of chosung strings
 	def chosungs
 		config['structure']['chosung']
 	end
 
-	# Array of jungsung's
+	# Array of jungsung's.
+	# @return [Array] Array of jungsung strings
 	def jungsungs
 		config['structure']['jungsung']
 	end
 
-	# Array of jongsung's
+	# Array of jongsung's.
+	# @return [Array] Array of jongsung strings
 	def jongsungs
 		config['structure']['jongsung']
 	end
 
-	# Checks if the given character is a korean character
+	# Checks if the given character is a korean character.
+	# @param [String] ch A string of size 1
 	def korean_char? ch
 		raise ArgumentError.new('Lengthy input') if ch.length > 1
 
@@ -43,6 +49,7 @@ class Korean
 
 	# Checks if the given character is a "complete" korean character.
 	# "Complete" Korean character must have chosung and jungsung, with optional jongsung.
+	# @param [String] ch A string of size 1
 	def complete_korean_char? ch
 		raise ArgumentError.new('Lengthy input') if ch.length > 1
 
@@ -50,14 +57,18 @@ class Korean
 		ch.unpack('U').all? { | c | c >= 0xAC00 && c <= 0xD7A3 }
 	end
 
-	# Splits the given string into an array of Korean::Char's and strings.
+	# Splits the given string into an array of Korean::Char's and Strings of length 1.
+	# @param [String] str Input string.
+	# @return [Array] Mixed array of Korean::Char instances and Strings of length 1 (for non-korean characters)
 	def dissect str
 		str.each_char.map { |c| 
 			korean_char?(c) ? Korean::Char.new(self, c) : c
 		}
 	end
 
-	# Reads a string with numbers in Korean way.
+	# Reads numeric expressions in Korean way.
+	# @param [String, Number] str Numeric type or String containing numeric expressions
+	# @return [String] Output string
 	def read_number str
 		nconfig = config['number']
 		
@@ -68,10 +79,13 @@ class Korean
 
 	# Returns the pronunciation of the given string containing Korean characters.
 	# Takes optional options hash.
-	# - If :pronounce_each_char is true, each character of the string is pronounced respectively.
-	# - If :slur is true, characters separated by whitespaces are treated as if they were contiguous.
-	# - If :number is true, numberic parts of the string is also pronounced in Korean.
-	# - :except array allows you to skip certain transformations.
+	#
+	# @param [String] Input string
+	# @param [Boolean] options[:pronounce_each_char] Each character of the string is pronounced respectively.
+	# @param [Boolean] options[:slur] Strings separated by whitespaces are processed again as if they were contiguous.
+	# @param [Boolean] options[:number] Numberic parts of the string is also pronounced in Korean.
+	# @param [Array] options[:except] Allows you to skip certain transformations.
+	# @return [String] Output string
 	def pronounce str, options = {}
 		options = {
 			:pronounce_each_char => false,
@@ -82,46 +96,24 @@ class Korean
 		}.merge options
 
 		str = read_number(str) if options[:number]
-		chars = dissect str
 
-		transforms = []
-		idx = -1
-		while (idx += 1) < chars.length
-			c = chars[idx]
-
-			next if c.is_a?(Korean::Char) == false
-
-			next_c = chars[idx + 1]
-			next_kc = (options[:pronounce_each_char] == false &&
-					   next_c.is_a?(Korean::Char) &&
-					   next_c.complete?) ? next_c : nil
-
-			transforms += @pronouncer.transform(c, next_kc, :except => options[:except])
-
-			# Slur (TBD)
-			if options[:slur] && options[:pronounce_each_char] == false && next_c =~ /\s/
-				chars[(idx + 1)..-1].each_with_index do | nc, new_idx |
-					next if nc =~ /\s/
-
-					if nc.is_a?(Korean::Char) && nc.complete?
-						transforms += @pronouncer.transform(c, nc, :except => options[:except])
-					end
-
-					idx = idx + 1 + new_idx - 1
-					break
-				end
-			end
-		end
+		result, transforms = @pronouncer.send :pronounce!, str, options
 
 		if options[:debug]
-			return chars.join, transforms
+			return result, transforms
 		else
-			chars.join
+			return result
 		end
 	end
 
 	# Returns the romanization (alphabetical notation) of the given Korean string.
 	# http://en.wikipedia.org/wiki/Korean_romanization
+	# @param [String] str Input Korean string
+	# @param [Boolean] options[:as_pronounced] If true, #pronounce is internally called before romanize
+	# @param [Boolean] options[:number] Whether to read numeric expressions in the string
+	# @param [Boolean] options[:slur] Same as :slur in #pronounce 
+	# @return [String] Output string in Roman Alphabet
+	# @see Korean#pronounce
 	def romanize str, options = {}
 		options = {
 			:as_pronounced => true,
@@ -142,23 +134,37 @@ class Korean
 				:except => %w[rule_5_3]
 		dash = rdata[0]["ㅇ"]
 		romanization = ""
-		(chars = str.each_char.to_a).each_with_index do | kc, cidx |
-			if korean_char? kc
-				Korean::Char.new(self, kc).to_a.each_with_index do | comp, idx |
+
+		romanize_chunk = lambda do | chunk |
+			dissect(chunk).each do | kc |
+				kc.to_a.each_with_index do | comp, idx |
 					next if comp.nil?
 					comp = rdata[idx][comp] || comp
 					comp = comp[1..-1] if comp[0] == dash &&
 							(romanization.empty? || romanization[-1] =~ /\s/ || comp[1] == 'w')
 					romanization += comp
 				end
-			else
-				romanization += kc
 			end
+
+			return post_subs.keys.inject(romanization) { | output, pattern |
+				output.gsub(pattern, post_subs[pattern])
+			}
 		end
 
-		post_subs.keys.inject(romanization) { | output, pattern |
-			output.gsub(pattern, post_subs[pattern])
-		}.capitalize
+		k_chunk = ""
+		str.each_char do | c |
+			if korean_char? c
+				k_chunk += c
+			else
+				unless k_chunk.empty?
+					romanization = romanize_chunk.call k_chunk
+					k_chunk = ""
+				end
+				romanization += c
+			end
+		end
+		romanization = romanize_chunk.call k_chunk unless k_chunk.empty?
+		romanization
 	end
 
 private

data/lib/gimchi/pronouncer.rb

@@ -2,47 +2,98 @@
 
 module Gimchi
 class Korean
-private
+	# Private class.
 	# Partial implementation of Korean pronouncement pronunciation rules specified in
 	# http://http://www.korean.go.kr/
 	class Pronouncer
-		attr_reader :applied
-
-		def initialize(korean)
+	private
+		def initialize korean
 			@korean = korean
 			@pconfig = korean.config['pronouncer']
-			@applied = []
 		end
 
-		def transform kc, next_kc, options = {}
-			options = { :except => [] }.merge options
-			@applied.clear
+		def pronounce! str, options = {}
+			@sequence = @pconfig['transformation']['sequence for ' +
+							(options[:pronounce_each_char] ? '1' : '2')] - options[:except]
 
-			# Cannot properly pronounce
-			return if kc.chosung.nil? && kc.jungsung.nil?
+			# Dissecting
+			@chars = @korean.dissect str
+			@orig_chars = @chars.dup
 
 			# Padding
-			kc.chosung = 'ㅇ' if kc.chosung.nil?
-			kc.jungsung = 'ㅡ' if kc.jungsung.nil?
+			@chars.each { |c| pad c }
 
-			if next_kc.nil?
-				rule_single kc, :except => options[:except]
-			else
-				not_todo = []
-				blocking_rule = @pconfig['transformation']['blocking rule']
-				@pconfig['transformation']['sequence'].each do | rule |
-					next if not_todo.include?(rule) || options[:except].include?(rule)
-
-					if self.send(rule, kc, next_kc)
-						@applied << rule
-						not_todo += blocking_rule[rule] if blocking_rule.has_key?(rule)
-					end
+			# Two-phase processing
+			# - For `slur'
+			applied = []
+			2.times do | phase |
+				@chars = @chars.reject { |c| c =~ /\s/ } if phase == 1
+
+				# Deep-fried...no copied backup
+				@initial_chars = @chars.map { |c| c.dup }
+
+				# Transform one by one
+				applied += (0...@chars.length).inject([]) { | arr, i | arr + transform(i); }
+
+				# Post-processing (actually just for :pronounce_each_char option)
+				@chars.select { |c| c.is_a?(Korean::Char) && c.jongsung }.each do | c |
+					c.jongsung = @pconfig['jongsung sound'][c.jongsung]
 				end
+
+				break unless options[:slur]
 			end
-			@applied
+
+			return @orig_chars.join, applied
 		end
 
 	private
+		def transform idx
+			@cursor = idx
+
+			# Not korean
+			return [] unless kc.is_a? Korean::Char
+
+			# Cannot properly pronounce
+			return [] if kc.chosung.nil? && kc.jungsung.nil? && kc.jongsung.nil?
+
+			applied = []
+			not_todo = []
+			blocking_rule = @pconfig['transformation']['blocking rule']
+			@sequence.each do | rule |
+				next if not_todo.include?(rule)
+
+				if self.send(rule,)
+					applied << rule
+					not_todo += blocking_rule[rule] if blocking_rule.has_key?(rule)
+				end
+			end
+			applied
+		end
+
+		def pad c
+			return unless c.is_a? Korean::Char
+
+			c.chosung = 'ㅇ'  if c.chosung.nil?
+			c.jungsung = 'ㅡ' if c.jungsung.nil?
+		end
+
+		def kc
+			@chars[@cursor]
+		end
+
+		def next_kc
+			nkc = @chars[@cursor + 1]
+			nkc.is_a?(Korean::Char) ? nkc : nil
+		end
+
+		def kc_org
+			@initial_chars[@cursor]
+		end
+
+		def next_kc_org
+			@initial_chars[@cursor + 1]
+		end
+
 		# shortcut
 		def fortis_map
 			@korean.config['structure']['fortis map']
@@ -53,20 +104,10 @@ private
 			@korean.config['structure']['double consonant map']
 		end
 
-		def rule_single kc, options = {}
-			options = {:except => []}.merge options
-			rule_5_1 kc, nil unless options[:except].include? 'rule_5_1'
-			rule_5_3 kc, nil unless options[:except].include? 'rule_5_3'
-
-			if kc.jongsung
-				kc.jongsung = @pconfig['jongsung sound'][kc.jongsung]
-			end
-		end
-
 		# 제5항: ‘ㅑ ㅒ ㅕ ㅖ ㅘ ㅙ ㅛ ㅝ ㅞ ㅠ ㅢ’는 이중 모음으로 발음한다.
 		#  다만 1. 용언의 활용형에 나타나는 ‘져, 쪄, 쳐’는 [저, 쩌, 처]로 발음한다.
 		#  다만 3. 자음을 첫소리로 가지고 있는 음절의 ‘ㅢ’는 [ㅣ]로 발음한다.
-		def rule_5_1 kc, next_kc
+		def rule_5_1
 			if %w[져 쪄 쳐].include? kc.to_s
 				kc.jungsung = 'ㅓ'
 
@@ -74,8 +115,8 @@ private
 			end
 		end
 
-		def rule_5_3 kc, next_kc
-			if kc.jungsung == 'ㅢ' && kc.org.chosung.consonant?
+		def rule_5_3
+			if kc.jungsung == 'ㅢ' && kc_org.chosung.consonant?
 				kc.jungsung = 'ㅣ'
 
 				true
@@ -84,7 +125,7 @@ private
 
 		# 제9항: 받침 ‘ㄲ, ㅋ’, ‘ㅅ, ㅆ, ㅈ, ㅊ, ㅌ’, ‘ㅍ’은 어말 또는 자음 앞에서
 		# 각각 대표음 [ㄱ, ㄷ, ㅂ]으로 발음한다.
-		def rule_9 kc, next_kc
+		def rule_9
 			map = {
 				%w[ㄲ ㅋ] => 'ㄱ',
 				%w[ㅅ ㅆ ㅈ ㅊ ㅌ] => 'ㄷ',
@@ -99,7 +140,7 @@ private
 
 		# 제10항: 겹받침 ‘ㄳ’, ‘ㄵ’, ‘ㄼ, ㄽ, ㄾ’, ‘ㅄ’은 어말 또는 자음 앞에서
 		# 각각 [ㄱ, ㄴ, ㄹ, ㅂ]으로 발음한다.
-		def rule_10 kc, next_kc
+		def rule_10
 			map = {
 				%w[ㄳ] => 'ㄱ',
 				%w[ㄵ] => 'ㄴ',
@@ -110,7 +151,7 @@ private
 				# Exceptions
 				if next_kc && (
 					   (kc.to_s == '밟' && next_kc.chosung.consonant?) ||
-					   (kc.to_s == '넓' && next_kc && %w[적 죽 둥].include?(next_kc.org.to_s))) # PATCH
+					   (kc.to_s == '넓' && next_kc && %w[적 죽 둥].include?(next_kc_org.to_s))) # PATCH
 					kc.jongsung = 'ㅂ'
 				else
 					kc.jongsung = map[ map.keys.find { |e| e.include? kc.jongsung } ]
@@ -121,7 +162,7 @@ private
 		end
 
 		# 제11항: 겹받침 ‘ㄺ, ㄻ, ㄿ’은 어말 또는 자음 앞에서 각각 [ㄱ, ㅁ, ㅂ]으로 발음한다.
-		def rule_11 kc, next_kc
+		def rule_11
 			map = {
 				'ㄺ' => 'ㄱ',
 				'ㄻ' => 'ㅁ',
@@ -131,7 +172,7 @@ private
 				# 다만, 용언의 어간 말음 ‘ㄺ’은 ‘ㄱ’ 앞에서 [ㄹ]로 발음한다.
 				# - 용언 여부 판단은?: 중성으로 판단 (PATCH)
 				if next_kc && kc.jongsung == 'ㄺ' &&
-					next_kc.org.chosung == 'ㄱ' &&
+					next_kc_org.chosung == 'ㄱ' &&
 					%w[맑 얽 섥 밝 늙 묽 넓].include?(kc.to_s) # PATCH
 					kc.jongsung = 'ㄹ'
 				else
@@ -155,7 +196,7 @@ private
 		#   [붙임]‘ㄶ, ㅀ’ 뒤에 ‘ㄴ’이 결합되는 경우에는, ‘ㅎ’을 발음하지 않는다.
 		#
 		#  4. ‘ㅎ(ㄶ, ㅀ)’ 뒤에 모음으로 시작된 어미나 접미사가 결합되는 경우에는, ‘ㅎ’을 발음하지 않는다.
-		def rule_12 kc, next_kc
+		def rule_12
 			return if next_kc.nil?
 
 			map_12_1 = {
@@ -218,17 +259,18 @@ private
 
 		# 제13항: 홑받침이나 쌍받침이 모음으로 시작된 조사나 어미, 접미사와
 		# 결합되는 경우에는, 제 음가대로 뒤 음절 첫소리로 옮겨 발음한다.
-		def rule_13 kc, next_kc
+		def rule_13
 			return if kc.jongsung.nil? || kc.jongsung == 'ㅇ' || next_kc.nil? || next_kc.chosung != 'ㅇ'
 			next_kc.chosung = kc.jongsung
 			kc.jongsung = nil
 
 			true
 		end
+
 		# 제14항: 겹받침이 모음으로 시작된 조사나 어미, 접미사와 결합되는 경우에는,
 		# 뒤엣것만을 뒤 음절 첫소리로 옮겨 발음한다.(이 경우, ‘ㅅ’은 된소리로 발음함.)
 		#
-		def rule_14 kc, next_kc
+		def rule_14
 			return if kc.jongsung.nil? || kc.jongsung == 'ㅇ' || next_kc.nil? || next_kc.chosung != 'ㅇ'
 			if consonants = double_consonant_map[kc.jongsung]
 				consonants[1] = 'ㅆ' if consonants[1] == 'ㅅ'
@@ -237,9 +279,10 @@ private
 				true
 			end
 		end
+
 		# 제15항: 받침 뒤에 모음 ‘ㅏ, ㅓ, ㅗ, ㅜ, ㅟ’들로 시작되는 __실질 형태소__가 연결되는
 		# 경우에는, 대표음으로 바꾸어서 뒤 음절 첫소리로 옮겨 발음한다.
-		def rule_15 kc, next_kc
+		def rule_15
 			return if kc.jongsung.nil? || kc.jongsung == 'ㅇ' || next_kc.nil? || next_kc.chosung != 'ㅇ'
 
 			if false && %w[ㅏ ㅓ ㅗ ㅜ ㅟ].include?(next_kc.jungsung) &&
@@ -253,7 +296,7 @@ private
 
 		# 제16항: 한글 자모의 이름은 그 받침소리를 연음하되, ‘ㄷ, ㅈ, ㅊ, ㅋ, ㅌ,
 		# ㅍ, ㅎ’의 경우에는 특별히 다음과 같이 발음한다.
-		def rule_16 kc, next_kc
+		def rule_16
 			return if next_kc.nil?
 
 			map = {'디귿' => '디긋',
@@ -278,7 +321,7 @@ private
 		# [ㅈ, ㅊ]으로 바꾸어서 뒤 음절 첫소리로 옮겨 발음한다.
 		#
 		# [붙임] ‘ㄷ’ 뒤에 접미사 ‘히’가 결합되어 ‘티’를 이루는 것은 [치]로 발음한다.
-		def rule_17 kc, next_kc
+		def rule_17
 			return if next_kc.nil? || %w[ㄷ ㅌ ㄾ].include?(kc.jongsung) == false
 
 			if next_kc.to_s == '이'
@@ -296,7 +339,7 @@ private
 
 		# 제18항: 받침 ‘ㄱ(ㄲ, ㅋ, ㄳ, ㄺ), ㄷ(ㅅ, ㅆ, ㅈ, ㅊ, ㅌ, ㅎ), ㅂ(ㅍ, ㄼ,
 		# ㄿ, ㅄ)’은 ‘ㄴ, ㅁ’ 앞에서 [ㅇ, ㄴ, ㅁ]으로 발음한다.
-		def rule_18 kc, next_kc
+		def rule_18
 			map = {
 				%w[ㄱ ㄲ ㅋ ㄳ ㄺ] => 'ㅇ', 
 				%w[ㄷ ㅅ ㅆ ㅈ ㅊ ㅌ ㅎ] => 'ㄴ', 
@@ -311,7 +354,7 @@ private
 
 		# 제19항: 받침 ‘ㅁ, ㅇ’ 뒤에 연결되는 ‘ㄹ’은 [ㄴ]으로 발음한다.
 		# [붙임]받침 ‘ㄱ, ㅂ’ 뒤에 연결되는 ‘ㄹ’도 [ㄴ]으로 발음한다.
-		def rule_19 kc, next_kc
+		def rule_19
 			if next_kc && next_kc.chosung == 'ㄹ' && %w[ㅁ ㅇ ㄱ ㅂ].include?(kc.jongsung)
 				next_kc.chosung = 'ㄴ'
 
@@ -325,11 +368,11 @@ private
 		end
 
 		# 제20항: ‘ㄴ’은 ‘ㄹ’의 앞이나 뒤에서 [ㄹ]로 발음한다.
-		def rule_20 kc, next_kc
+		def rule_20
 			return if next_kc.nil?
 
 			to = if %w[견란 진란 산량 단력 권력 원령 견례
-						문로 단로 원론 원료 근류].include?(kc.org.to_s + next_kc.org.to_s)
+						문로 단로 원론 원료 근류].include?(kc_org.to_s + next_kc_org.to_s)
 					 'ㄴ'
 				 else
 					 'ㄹ'
@@ -348,7 +391,7 @@ private
 
 		# 제23항: 받침 ‘ㄱ(ㄲ, ㅋ, ㄳ, ㄺ), ㄷ(ㅅ, ㅆ, ㅈ, ㅊ, ㅌ), ㅂ(ㅍ, ㄼ, ㄿ,ㅄ)’
 		# 뒤에 연결되는 ‘ㄱ, ㄷ, ㅂ, ㅅ, ㅈ’은 된소리로 발음한다.
-		def rule_23 kc, next_kc
+		def rule_23
 			return if next_kc.nil?
 			if fortis_map.keys.include?(next_kc.chosung) &&
 				%w[ㄱ ㄲ ㅋ ㄳ ㄺ ㄷ ㅅ ㅆ ㅈ ㅊ ㅌ ㅂ ㅍ ㄼ ㄿ ㅄ].include?(kc.jongsung)
@@ -361,7 +404,7 @@ private
 		# 제24항: 어간 받침 ‘ㄴ(ㄵ), ㅁ(ㄻ)’ 뒤에 결합되는 어미의 첫소리 ‘ㄱ, ㄷ, ㅅ, ㅈ’은 된소리로 발음한다.
 		# 다만, 피동, 사동의 접미사 ‘-기-’는 된소리로 발음하지 않는다.
 		# 용언 어간에만 적용.
-		def rule_24 kc, next_kc
+		def rule_24
 			return if next_kc.nil? || 
 				next_kc.to_s == '기' # FIXME 피동/사동 여부 판단 불가. e.g. 줄넘기
 
@@ -385,7 +428,7 @@ private
 
 		# 제25항: 어간 받침 ‘ㄼ, ㄾ’ 뒤에 결합되는 어미의 첫소리 ‘ㄱ, ㄷ, ㅅ, ㅈ’은
 		# 된소리로 발음한다.
-		def rule_25 kc, next_kc
+		def rule_25
 			return if next_kc.nil?
 
 			if %w[ㄱ ㄷ ㅅ ㅈ].include?(next_kc.chosung) &&
@@ -397,13 +440,13 @@ private
 		end
 
 		# 제26항: 한자어에서, ‘ㄹ’ 받침 뒤에 연결되는 ‘ㄷ, ㅅ, ㅈ’은 된소리로 발음한다.
-		def rule_26 kc, next_kc
+		def rule_26
 			# TODO
 		end
 
 		# 제27항: __관형사형__ ‘-(으)ㄹ’ 뒤에 연결되는 ‘ㄱ, ㄷ, ㅂ, ㅅ, ㅈ’은 된소리로 발음한다.
 		# - ‘-(으)ㄹ’로 시작되는 어미의 경우에도 이에 준한다.
-		def rule_27 kc, next_kc
+		def rule_27
 			# FIXME: NOT PROPERLY IMPLEMENTED
 			return if next_kc.nil?
 
@@ -419,14 +462,14 @@ private
 		# 제28항: 표기상으로는 사이시옷이 없더라도, 관형격 기능을 지니는 사이시옷이
 		# 있어야 할(휴지가 성립되는) 합성어의 경우에는, 뒤 단어의 첫소리 ‘ㄱ, ㄷ,
 		# ㅂ, ㅅ, ㅈ’을 된소리로 발음한다.
-		def rule_26_28 kc, next_kc
+		def rule_26_28
 			# TODO
 		end
 
 		# 제29항: 합성어 및 파생어에서, 앞 단어나 접두사의 끝이 자음이고 뒤 단어나
 		# 접미사의 첫음절이 ‘이, 야, 여, 요, 유’인 경우에는, ‘ㄴ’ 음을 첨가하여
 		# [니, 냐, 녀, 뇨, 뉴]로 발음한다.
-		def rule_29 kc, next_kc
+		def rule_29
 			# TODO
 		end
 
@@ -436,7 +479,7 @@ private
 		# 발음하는 것도 허용한다.
 		# 2. 사이시옷 뒤에 ‘ㄴ, ㅁ’이 결합되는 경우에는 [ㄴ]으로 발음한다. 
 		# 3. 사이시옷 뒤에 ‘이’ 음이 결합되는 경우에는 [ㄴㄴ]으로 발음한다.
-		def rule_30 kc, next_kc
+		def rule_30
 			return if next_kc.nil? || kc.jongsung != 'ㅅ'
 
 			if %w[ㄱ ㄷ ㅂ ㅅ ㅈ].include? next_kc.chosung

data/test/romanization.yml

@@ -1,4 +1,6 @@
 ---
+"반기문": "Ban-gimun"
+"방이문": "Bang-imun"
 "구미": "Gumi"
 "영동": "Yeongdong"
 "백암": "Baegam"

data/test/test_gimchi.rb

@@ -97,18 +97,30 @@ class TestGimchi < Test::Unit::TestCase
 		test_set.each do | k, v |
 			cnt += 1
 			k = k.gsub(/[-]/, '')
-			t, tfs = ko.pronounce(k, :pronounce_each_char => false, :slur => k.include?(' '), :debug => true)
-			if v.include? t.gsub(/\s/, '')
+
+			t1, tfs1 = ko.pronounce(k, :pronounce_each_char => false, :slur => true, :debug => true)
+			t2, tfs2 = ko.pronounce(k, :pronounce_each_char => false, :slur => false, :debug => true)
+
+			path = ""
+			if (with_slur = v.include?(t1.gsub(/\s/, ''))) || v.include?(t2.gsub(/\s/, ''))
 				r = ANSI::Code::BLUE + ANSI::Code::BOLD + v.join(' / ') + ANSI::Code::RESET if v.length > 1
+				path = (with_slur ? tfs1 : tfs2).map { |e| e.sub 'rule_', '' }.join(' > ')
+				t = with_slur ? t1 : t2
 				s += 1
 			else
 				r = ANSI::Code::RED + ANSI::Code::BOLD + v.join(' / ') + ANSI::Code::RESET
+				t = [t1, t2].join ' | '
 			end
-			puts "#{k} => #{t} (#{ko.romanize t}) [#{tfs.join(' > ')}] #{r}"
+			puts "#{k} => #{t} (#{ko.romanize t, :as_pronounced => false}) [#{path}] #{r}"
 		end
 		puts "#{s} / #{cnt}"
 		# FIXME
-		assert s >= 410
+		assert s >= 411
+	end
+
+	def test_romanize_preservce_non_korean
+		ko = Gimchi::Korean.new
+		assert_equal 'ttok-kkateun kkk', ko.romanize('똑같은 kkk')
 	end
 
 	def test_romanize
@@ -130,6 +142,6 @@ class TestGimchi < Test::Unit::TestCase
 		end
 		puts "#{s} / #{cnt}"
 		# FIXME
-		assert s >= 55
+		assert s >= 57
 	end
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: gimchi
 version: !ruby/object:Gem::Version
-  version: 0.1.0
+  version: 0.1.1
   prerelease: 
 platform: ruby
 authors:
@@ -9,12 +9,11 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2011-04-05 00:00:00.000000000 +09:00
-default_executable: 
+date: 2011-04-07 00:00:00.000000000Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
-  requirement: &2156236300 !ruby/object:Gem::Requirement
+  requirement: &2153047260 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
@@ -22,10 +21,10 @@ dependencies:
         version: 1.0.0
   type: :development
   prerelease: false
-  version_requirements: *2156236300
+  version_requirements: *2153047260
 - !ruby/object:Gem::Dependency
   name: jeweler
-  requirement: &2156235820 !ruby/object:Gem::Requirement
+  requirement: &2153046780 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
@@ -33,10 +32,10 @@ dependencies:
         version: 1.5.2
   type: :development
   prerelease: false
-  version_requirements: *2156235820
+  version_requirements: *2153046780
 - !ruby/object:Gem::Dependency
   name: rcov
-  requirement: &2156235340 !ruby/object:Gem::Requirement
+  requirement: &2153046300 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -44,10 +43,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *2156235340
+  version_requirements: *2153046300
 - !ruby/object:Gem::Dependency
   name: ansi
-  requirement: &2156234860 !ruby/object:Gem::Requirement
+  requirement: &2153045820 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -55,8 +54,8 @@ dependencies:
         version: 1.2.2
   type: :development
   prerelease: false
-  version_requirements: *2156234860
-description: Gimchi knows how to pronounce Korean string and how to write them in
+  version_requirements: *2153045820
+description: Gimchi knows how to pronounce Korean strings and how to write them in
   roman alphabet.
 email: junegunn.c@gmail.com
 executables: []
@@ -78,7 +77,6 @@ files:
 - test/pronunciation.yml
 - test/romanization.yml
 - test/test_gimchi.rb
-has_rdoc: true
 homepage: http://github.com/junegunn/gimchi
 licenses:
 - MIT
@@ -100,7 +98,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
 requirements: []
 rubyforge_project: 
-rubygems_version: 1.6.2
+rubygems_version: 1.7.2
 signing_key: 
 specification_version: 3
 summary: Gimchi reads Korean.
@@ -109,3 +107,4 @@ test_files:
 - test/pronunciation.yml
 - test/romanization.yml
 - test/test_gimchi.rb
+has_rdoc: