gimchi 0.1.4 → 0.1.5

data/config/default.yml

@@ -84,14 +84,22 @@ pronouncer:
             rule_16: [rule_30]
 
 number:
+    positive: 플러스
     negative: 마이너스
     decimal point: 점
     units: ["", 만, 억, 조, 경, 해, 자, 양, 구, 간, 정, 재, 극, 항하사, 아승기, 나유타, 불가사의, 무량대수]
     digits: [영, 일, 이, 삼, 사, 오, 육, 칠, 팔, 구]
+    post substitution:
+        ? !ruby/regexp /^일만/
+        : 만
 
     # 정수형일 때 또다른 표현법 (나이, 시간)
     alt notation:
         when suffix:
+            개:
+                max: 
+            명:
+                max: 
             살:
                 max: 
             시:

data/lib/gimchi/korean.rb

@@ -72,8 +72,8 @@ class Korean
 	def read_number str
 		nconfig = config['number']
 		
-		str.to_s.gsub(/([+-]\s*)?[0-9,]*,*[0-9]+(\.[0-9]+)?(\s*.)?/) { 
-			read_number_sub($&, $3)
+		str.to_s.gsub(/([+-]\s*)?[0-9,]*,*[0-9]+(\.[0-9]+(e[+-][0-9]+)?)?(\s*.)?/) { 
+			read_number_sub($&, $4)
 		}
 	end
 
@@ -171,101 +171,120 @@ private
 	def read_number_sub num, next_char = nil
 		nconfig = config['number']
 
-		# To number
-		if num.is_a? String
-			num = num.gsub(/[\s,]/, '')
-			raise ArgumentError.new("Invalid number format") unless num =~ /[-+]?[0-9,]*\.?[0-9]*/
-			num = num.to_f == num.to_i ? num.to_i : num.to_f
-		end
+		num = num.gsub(',', '')
+		num = num.sub(/#{next_char}$/, '') if next_char
+		is_float = num.match(/[\.e]/) != nil
 
 		# Alternative notation for integers with proper suffix
 		alt = false
-		if num.is_a?(Float) == false && nconfig['alt notation']['when suffix'].keys.include?(next_char.to_s.strip)
+		if is_float == false && 
+				nconfig['alt notation']['when suffix'].keys.include?(next_char.to_s.strip)
 			max = nconfig['alt notation']['when suffix'][next_char.strip]['max']
 
-			if max.nil? || num <= max
+			if max.nil? || num.to_i <= max
 				alt = true
 			end
 		end
 
 		# Sign
-		if num < 0
-			num = -1 * num
+		sign = []
+		negative = false
+		if num =~ /^-/
+			num = num.sub(/^-\s*/, '')
+			sign << nconfig['negative']
 			negative = true
-		else
-			negative = false
+		elsif num =~ /^\+/
+			num = num.sub(/^\+\s*/, '')
+			sign << nconfig['positive']
 		end
 
-		if num.is_a? Float
+		if is_float
 			below = nconfig['decimal point']
-			below = nconfig['digits'][0] + below if num < 1
+			below = nconfig['digits'][0] + below if num.to_f < 1
 
-			s = num.to_s
-			if md = s.match(/(.*)e(.*)/)
-				s = md[1].tr '.', ''
+			if md = num.match(/(.*)e(.*)/)
+				dp = md[1].index('.')
+				num = md[1].tr '.', ''
 				exp = md[2].to_i
-				if exp > 0
-					s = s.ljust(exp + 1, '0')
+
+				dp += exp
+				if dp > num.length
+					num = num.ljust(dp, '0')
+					num = num.sub(/^0+([1-9])/, "\\1")
+
+					below = ""
+				elsif dp < 0
+					num = '0.' + '0' * (-dp) + num
 				else
-					s = '0.' + '0' * (-exp - 1) + s
+					num[dp] = '.' + num[dp]
 				end
 			end
-			s.sub(/.*\./, '').each_char do | char |
+			num.sub(/.*\./, '').each_char do | char |
 				below += nconfig['digits'][char.to_i]
-			end
-			num = num.floor.to_i
+			end if num.include? '.'
+			num = num.sub(/\..*/, '')
 		else
 			below = ""
 		end
 
 		tokens = []
 		unit_idx = -1
+		num = num.to_i
 		while num > 0
 			v = num % 10000
 
-			if alt == false || unit_idx >= 0
-				str = ""
-				# Cannot use hash as they're unordered in 1.8
-				[[1000, '천'],
-				 [100,  '백'],
-				 [10,   '십']].each do | arr |
-				 	u, sub_unit = arr
-					str += (nconfig['digits'][v/u] if v/u != 1).to_s + sub_unit + ' ' if v / u > 0
-					v %= u
-				end
-				str += nconfig['digits'][v] if v > 0
-
-				tokens << str.sub(/ $/, '') + nconfig['units'][unit_idx += 1]
-			else
-				str = ""
-				tenfolds = nconfig['alt notation']['tenfolds']
-				digits = nconfig['alt notation']['digits']
-				post_subs = nconfig['alt notation']['post substitution']
-
-				# Likewise.
-				[[1000, '천'],
-				 [100,  '백']].each do | u, sub_unit |
-					str += (nconfig['digits'][v/u] if v/u != 1).to_s + sub_unit + ' ' if v / u > 0
-					v %= u
-				end
-
-				str += tenfolds[(v / 10) - 1] if v / 10 > 0
-				v %= 10
-				str += digits[v] if v > 0
-
-				suffix = next_char.strip
-				str = str + suffix
-				post_subs.each do | k, v |
-					str.gsub!(k, v)
+			unit_idx += 1
+			if v > 0
+				if alt == false || unit_idx >= 1
+					str = ""
+					# Cannot use hash as they're unordered in 1.8
+					[[1000, '천'],
+					 [100,  '백'],
+					 [10,   '십']].each do | arr |
+						u, sub_unit = arr
+						str += (nconfig['digits'][v/u] if v/u != 1).to_s + sub_unit + ' ' if v / u > 0
+						v %= u
+					end
+					str += nconfig['digits'][v] if v > 0
+
+					tokens << str.sub(/ $/, '') + nconfig['units'][unit_idx]
+				else
+					str = ""
+					tenfolds = nconfig['alt notation']['tenfolds']
+					digits = nconfig['alt notation']['digits']
+					alt_post_subs = nconfig['alt notation']['post substitution']
+
+					# Likewise.
+					[[1000, '천'],
+					 [100,  '백']].each do | u, sub_unit |
+						str += (nconfig['digits'][v/u] if v/u != 1).to_s + sub_unit + ' ' if v / u > 0
+						v %= u
+					end
+
+					str += tenfolds[(v / 10) - 1] if v / 10 > 0
+					v %= 10
+					str += digits[v] if v > 0
+
+					if alt
+						suffix = next_char.strip
+						str = str + suffix
+						alt_post_subs.each do | k, v |
+							str.gsub!(k, v)
+						end
+						str.sub!(/#{suffix}$/, '')
+					end
+					tokens << str.sub(/ $/, '') + nconfig['units'][unit_idx]
 				end
-				str.sub!(/#{suffix}$/, '')
-				tokens << str.sub(/ $/, '') + nconfig['units'][unit_idx += 1]
 			end
 			num /= 10000
 		end
 
-		tokens << nconfig['negative'] if negative
-		tokens.reverse.join(' ') + next_char.to_s + below
+		tokens += sign unless sign.empty?
+		ret = tokens.reverse.join(' ') + below + next_char.to_s
+		nconfig['post substitution'].each do | k, v |
+			ret.gsub!(k, v)
+		end
+		ret
 	end
 end#Korean
 end#Gimchi

data/test/test_gimchi.rb

@@ -77,12 +77,25 @@ class TestGimchi < Test::Unit::TestCase
 		ko = Gimchi::Korean.new
 		assert_equal "천 구백 구십 구", ko.read_number(1999)
 		assert_equal "마이너스 백점일이삼", ko.read_number(- 100.123)
-		assert_equal "오백 삼십 일억 구천 백 십만 육백 칠십 팔점삼이일사",
-				ko.read_number("53,191,100,678.3214")
-
-		# 나이, 시간 ( -살, -시 )
+		assert_equal "오백 삼십 일억 구천 백 십만 육백 칠십 팔점삼이일사오육칠",
+				ko.read_number("53,191,100,678.3214567")
+		assert_equal "영점영영영영영일이삼사오", ko.read_number("1.2345e-06")
+		assert_equal "일해 이천 삼백 사십 오경", ko.read_number("1.2345e+20")
+		assert_equal "플러스 일해 이천 삼백 사십 오경", ko.read_number("+ 1.2345e+20")
+		assert_equal "마이너스 일해 이천 삼백 사십 오경", ko.read_number("- 1.2345e+20")
+		assert_equal "만 십 이점삼", ko.read_number("100.123e+2")
+		assert_equal "십만 십 이점삼", ko.read_number("1000.123e+2")
+		assert_equal "백 일만 십 이점삼", ko.read_number("10100.123e+2")
+		assert_equal "천 십 이점삼", ko.read_number("10.123e+2")
+		assert_equal "십점영", ko.read_number("10.0")
+		assert_equal "플러스 십점영", ko.read_number("+ 10.0")
+
+		# 나이, 시간, 개수, 명 ( -살, -시, -개, -명 )
 		assert_equal "나는 스무살", ko.read_number("나는 20살")
+		assert_equal "이십 칠점일살", ko.read_number("27.1살")
 		assert_equal "너는 열세 살", ko.read_number("너는 13 살")
+		assert_equal "백 서른두명", ko.read_number("132명")
+		assert_equal "이천 오백 아흔아홉개", ko.read_number("2,599개")
 		assert_equal "지금은 일곱시 삼십분", ko.read_number("지금은 7시 30분")
 	end
 
@@ -118,7 +131,7 @@ class TestGimchi < Test::Unit::TestCase
 		assert s >= 411
 	end
 
-	def test_romanize_preservce_non_korean
+	def test_romanize_preserve_non_korean
 		ko = Gimchi::Korean.new
 		assert_equal 'ttok-kkateun kkk', ko.romanize('똑같은 kkk')
 	end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: gimchi
 version: !ruby/object:Gem::Version
-  version: 0.1.4
+  version: 0.1.5
   prerelease: 
 platform: ruby
 authors:
@@ -9,11 +9,11 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2011-04-08 00:00:00.000000000Z
+date: 2011-04-12 00:00:00.000000000Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
-  requirement: &2152726760 !ruby/object:Gem::Requirement
+  requirement: &2152493140 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
@@ -21,10 +21,10 @@ dependencies:
         version: 1.0.0
   type: :development
   prerelease: false
-  version_requirements: *2152726760
+  version_requirements: *2152493140
 - !ruby/object:Gem::Dependency
   name: jeweler
-  requirement: &2152726280 !ruby/object:Gem::Requirement
+  requirement: &2152491800 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
@@ -32,10 +32,10 @@ dependencies:
         version: 1.5.2
   type: :development
   prerelease: false
-  version_requirements: *2152726280
+  version_requirements: *2152491800
 - !ruby/object:Gem::Dependency
   name: rcov
-  requirement: &2152742180 !ruby/object:Gem::Requirement
+  requirement: &2152490620 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -43,10 +43,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *2152742180
+  version_requirements: *2152490620
 - !ruby/object:Gem::Dependency
   name: ansi
-  requirement: &2152741700 !ruby/object:Gem::Requirement
+  requirement: &2152489680 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -54,7 +54,7 @@ dependencies:
         version: 1.2.2
   type: :development
   prerelease: false
-  version_requirements: *2152741700
+  version_requirements: *2152489680
 description: Gimchi knows how to pronounce Korean strings and how to write them in
   roman alphabet.
 email: junegunn.c@gmail.com
@@ -91,6 +91,9 @@ required_ruby_version: !ruby/object:Gem::Requirement
   - - ! '>='
     - !ruby/object:Gem::Version
       version: '0'
+      segments:
+      - 0
+      hash: -4096346844211308775
 required_rubygems_version: !ruby/object:Gem::Requirement
   none: false
   requirements: