zhconv 0.1.1

data/lib/zhconv.rb

@@ -0,0 +1,96 @@
+# encoding: utf-8
+path = File.expand_path(File.dirname(__FILE__))
+$:.unshift(path) unless $:.include?(path)
+
+require 'open-uri'
+
+module ZhConv
+  @converters = {}
+  VARIANTS  = ["zh-hans", "zh-hant", "zh-cn", "zh-sg", "zh-tw", "zh-hk"]
+
+  class Converter
+    def initialize(url)
+      @mapping = {}    
+      table = open(url).read
+      table.lines do |line|
+        matches = line.match(/^[\*\"]([^\"=]+)\"?\s*=>\s*\"?([^\s\/\"]+)\s?.*\"?[;,]$/)
+        if matches
+          @mapping[matches[1].strip] = matches[2].strip
+        end
+      end
+      #puts "create converter for (#{url}), words: #{@mapping.size}"
+    end
+    
+    def convert(input)
+      @mapping.each_key do |key|
+        input.gsub!(key, @mapping[key])
+      end
+      input
+    end
+  end
+
+  def self.convert(variant, message, use_web=true)    
+    if use_web
+      case variant
+      when "zh-cn", "zh-sg"
+        converter = web_converter(variant)
+        message = converter.convert(message)
+      when "zh-tw", "zh-hk"
+        converter = web_converter(variant)
+        message = converter.convert(message)
+      end
+    end
+    
+    case variant
+    when "zh-cn", "zh-sg"
+      converter = local_converter(variant)
+      message = converter.convert(message)
+    when "zh-tw", "zh-hk"
+      converter = local_converter(variant)
+      message = converter.convert(message)
+    end
+    
+    if use_web  
+      case variant
+      when "zh-hans", "zh-cn", "zh-sg"
+        converter = web_converter("zh-hans")
+        message = converter.convert(message)
+      when "zh-hant", "zh-tw", "zh-hk"
+        converter = web_converter("zh-hant")
+        message = converter.convert(message)
+      end
+    end
+    
+    case variant
+    when "zh-hans", "zh-cn", "zh-sg"
+      converter = local_converter("zh-hans")
+      message = converter.convert(message)
+    when "zh-hant", "zh-tw", "zh-hk"
+      converter = local_converter("zh-hant")
+      message = converter.convert(message)
+    end
+
+    message
+  end
+
+  def self.local_converter(variant)
+    raise "variant #{variant} not supported" unless VARIANTS.index(variant)
+    url = converter_url(variant, false)   
+    @converters[url] ||= Converter.new(url)
+  end
+
+  def self.web_converter(variant)
+    raise "variant #{variant} not supported" unless VARIANTS.index(variant)
+    url = converter_url(variant, true)    
+    @converters[url] ||= Converter.new(url)
+  end
+  
+  def self.converter_url(variant, web_converter=false)
+    if web_converter
+      "http://zh.wikipedia.org/w/index.php?title=MediaWiki:Conversiontable/#{variant}&action=raw&templates=expand"
+    else
+      File.expand_path(File.dirname(__FILE__)) + "/../data/#{variant}.txt"
+    end
+  end
+  
+end

data/test/test_zhconv.rb

@@ -0,0 +1,64 @@
+# encoding: utf-8
+
+require 'test/unit'
+require "zhconv"
+
+class TestZhConv < Test::Unit::TestCase
+  def test_conv_trad    
+    assert_equal "歷史",    ZhConv.convert("zh-hant", "历史", false)
+    assert_equal "麵包",    ZhConv.convert("zh-hant", "面包", false)
+    assert_equal "獅子山",    ZhConv.convert("zh-hant", "狮子山", false)
+    assert_equal "五隻",    ZhConv.convert("zh-hant", "五只", false)
+    
+    text_zh_tw = "蜻蜓，是一種屬於蜻蛉目的昆蟲，於春天到秋天時發生，有細長翅膀與腹部。"
+    text_zh_cn = "蜻蜓，是一种属于蜻蛉目的昆虫，于春天到秋天时发生，有细长翅膀与腹部。"
+    assert_equal text_zh_tw, ZhConv.convert("zh-hant", text_zh_cn, false)
+  end
+
+  def test_conv_hk
+    assert_equal "歷史",    ZhConv.convert("zh-hk", "历史", false)
+    assert_equal "麵包",    ZhConv.convert("zh-hk", "面包", false)
+    assert_equal "獅子山",    ZhConv.convert("zh-hk", "狮子山", false)
+    assert_equal "羅納爾多",    ZhConv.convert("zh-hk", "罗纳尔多", false)
+  end
+  
+  def test_conv_simp
+    assert_equal "历史",    ZhConv.convert("zh-hans", "曆史", false)
+    assert_equal "面包",    ZhConv.convert("zh-hans", "麵包", false)
+    assert_equal "远程控制",    ZhConv.convert("zh-hans", "遠程控制", false)
+
+    text_zh_tw = "蜻蜓，是一種屬於蜻蛉目的昆蟲，於春天到秋天時發生，有細長翅膀與腹部。"
+    text_zh_cn = "蜻蜓，是一种属于蜻蛉目的昆虫，于春天到秋天时发生，有细长翅膀与腹部。"
+    assert_equal text_zh_cn, ZhConv.convert("zh-hans", text_zh_tw, false)
+  end
+  
+  def test_text
+    text_zh_tw = "蜻蜓，是一種屬於蜻蛉目的昆蟲，於春天到秋天時發生，有細長翅膀與腹部。"
+    text_zh_cn = "蜻蜓，是一种属于蜻蛉目的昆虫，于春天到秋天时发生，有细长翅膀与腹部。"
+    assert_equal text_zh_cn, ZhConv.convert("zh-cn", text_zh_tw, false)
+  end
+
+
+  def test_conv_trad_web
+    assert_equal "五隻",    ZhConv.convert("zh-hant", "五只")
+  end
+  
+  def test_conv_hk_web
+    assert_equal "朗拿度",    ZhConv.convert("zh-hk", "罗纳尔多")
+  end
+  
+  def test_conv_simp_web
+    assert_equal "远程控制",    ZhConv.convert("zh-hans", "遠程控制")  
+  end
+  
+  def test_conv_cn_web
+    assert_equal "冰激凌",    ZhConv.convert("zh-cn", "冰淇淋")
+    assert_equal "东加拿大",    ZhConv.convert("zh-cn", "東加拿大")
+  end
+  
+  def test_conv_sg_web
+    assert_equal "健力士世界纪录",    ZhConv.convert("zh-sg", "吉尼斯世界纪录")
+    assert_equal "纽西兰",    ZhConv.convert("zh-sg", "新西蘭")
+  end
+
+end

data/zhconv.gemspec

@@ -0,0 +1,30 @@
+# -*- encoding: utf-8 -*-
+
+Gem::Specification.new do |s|
+  s.name = "zhconv"
+  s.version = "0.1.1"
+
+  s.required_rubygems_version = Gem::Requirement.new(">= 1.2") if s.respond_to? :required_rubygems_version=
+  s.authors = ["Francis Chong"]
+  s.date = "2011-11-16"
+  s.description = "Convert chinese using mediawiki conversion table"
+  s.email = "francis@ignition.hk"
+  s.extra_rdoc_files = ["README.md", "lib/zhconv.rb"]
+  s.files = ["Gemfile", "Gemfile.lock", "Manifest", "README.md", "Rakefile", "data/zh-cn.txt", "data/zh-hans.txt", "data/zh-hant.txt", "data/zh-hk.txt", "data/zh-sg.txt", "data/zh-tw.txt", "lib/zhconv.rb", "test/test_zhconv.rb", "zhconv.gemspec"]
+  s.homepage = "http://github.com/siuying/zhconv"
+  s.rdoc_options = ["--line-numbers", "--inline-source", "--title", "Zhconv", "--main", "README.md"]
+  s.require_paths = ["lib"]
+  s.rubyforge_project = "zhconv"
+  s.rubygems_version = "1.8.10"
+  s.summary = "Convert chinese using mediawiki conversion table"
+  s.test_files = ["test/test_zhconv.rb"]
+
+  if s.respond_to? :specification_version then
+    s.specification_version = 3
+
+    if Gem::Version.new(Gem::VERSION) >= Gem::Version.new('1.2.0') then
+    else
+    end
+  else
+  end
+end

metadata

@@ -0,0 +1,67 @@
+--- !ruby/object:Gem::Specification
+name: zhconv
+version: !ruby/object:Gem::Version
+  version: 0.1.1
+  prerelease: 
+platform: ruby
+authors:
+- Francis Chong
+autorequire: 
+bindir: bin
+cert_chain: []
+date: 2011-11-16 00:00:00.000000000 Z
+dependencies: []
+description: Convert chinese using mediawiki conversion table
+email: francis@ignition.hk
+executables: []
+extensions: []
+extra_rdoc_files:
+- README.md
+- lib/zhconv.rb
+files:
+- Gemfile
+- Gemfile.lock
+- Manifest
+- README.md
+- Rakefile
+- data/zh-cn.txt
+- data/zh-hans.txt
+- data/zh-hant.txt
+- data/zh-hk.txt
+- data/zh-sg.txt
+- data/zh-tw.txt
+- lib/zhconv.rb
+- test/test_zhconv.rb
+- zhconv.gemspec
+homepage: http://github.com/siuying/zhconv
+licenses: []
+post_install_message: 
+rdoc_options:
+- --line-numbers
+- --inline-source
+- --title
+- Zhconv
+- --main
+- README.md
+require_paths:
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement
+  none: false
+  requirements:
+  - - ! '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+required_rubygems_version: !ruby/object:Gem::Requirement
+  none: false
+  requirements:
+  - - ! '>='
+    - !ruby/object:Gem::Version
+      version: '1.2'
+requirements: []
+rubyforge_project: zhconv
+rubygems_version: 1.8.10
+signing_key: 
+specification_version: 3
+summary: Convert chinese using mediawiki conversion table
+test_files:
+- test/test_zhconv.rb