konjak 0.0.1 → 0.0.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 620fb76e02f4a35a2e085fb25e50d2cf5924d4a3
-  data.tar.gz: 13a5d704eceda0ab92f7bc73a7af9f42024f5fcb
+  metadata.gz: 3d1f7b71cef0bca6b16b61cccb6dae7fe49c28fc
+  data.tar.gz: 58e133fc23d9789b8075e4b1d0b497259193d436
 SHA512:
-  metadata.gz: f792f56294bcd16a38a8f71ad876d3fef54d5f1b00b53a93eb21f6456a7abce9229c541b8e16d81296d2bf4c580861fccc9d17c39421b418864c13aab50f90ce
-  data.tar.gz: 45cbd09c3b26c03d9f8236a9ea31bfc0d3116d1a54a124d7dbb7c82ab5ababfcae58e3b4556c17d884b13c315ac2e3e23bbbef070dbfadc921edc715fcf614b3
+  metadata.gz: 961ec5b4b2635480938016110b07b654b112353fe8c84dd3260bb812d2e06c9163c69dd957686b65ba619f3b2b634420850a203e8be41a070bd441fe55c4b884
+  data.tar.gz: 0d9385eda3e902e2519f5b34d7d0a9af8397859b70d8d2a687c9b75536a8904cbea68d0a7aaa10b8528baeb9b7655902259ea6d6e94a6da5c990937df2903945

data/.travis.yml

@@ -1,3 +1,6 @@
 language: ruby
 rvm:
-  - 2.2
+- 2.2
+env:
+  global:
+    secure: V8yG0G03Dsabc5mK4ubx3wSd0FyW0tazC/4DsJrvQTXSNytab1tXpwQbaN1XUyWTGnOj58mSZAPk/36n/5sGEwq3HIYCvPbJrqrCRODTDzmYy5uA1N+/Mo2uA55lPBjCbfcEvO1oy9tQIcH3I3R29nqiTsIRXax8XoFXHlJdP7I=

data/Gemfile

@@ -2,3 +2,7 @@ source 'https://rubygems.org'
 
 # Specify your gem's dependencies in konjak.gemspec
 gemspec
+
+if ENV['CODECLIMATE_REPO_TOKEN']
+  gem "codeclimate-test-reporter", group: :test, require: nil
+end

data/README.md

@@ -1,5 +1,9 @@
 # Konjak
 
+[![Build Status](https://travis-ci.org/hanachin/konjak.svg)](https://travis-ci.org/hanachin/konjak)
+[![Code Climate](https://codeclimate.com/github/hanachin/konjak/badges/gpa.svg)](https://codeclimate.com/github/hanachin/konjak)
+[![Test Coverage](https://codeclimate.com/github/hanachin/konjak/badges/coverage.svg)](https://codeclimate.com/github/hanachin/konjak/coverage)
+
 TMX(Translation Memory eXchange) tools for ruby
 
 ## Installation
@@ -18,7 +22,7 @@ Or install it yourself as:
 
 ## Usage
 
-    $ konjak translate file.tmx file.txt src target
+    $ konjak translate src target file.tmx file.txt
 
 ## Contributing
 

data/Rakefile

@@ -1 +1,11 @@
 require "bundler/gem_tasks"
+
+begin
+  require 'rspec/core/rake_task'
+
+  RSpec::Core::RakeTask.new(:spec)
+
+  task :default => :spec
+rescue LoadError
+  # no rspec available
+end

data/bin/konjak

@@ -13,4 +13,4 @@ _command, src, target, tmx_path, target_path = ARGV
 
 tmx = File.read(tmx_path)
 doc = File.read(target_path)
-puts Konjak.translate(doc, Konjak.parse(tmx, gtt: true), src, target)
+puts Konjak.translate(doc, Konjak.parse(tmx, gtt: true), src, target, format: :gtt_html)

data/lib/konjak/text.rb

@@ -1,15 +1,4 @@
 module Konjak
-  class Text
-    def initialize(text)
-      @text = text
-    end
-
-    def to_s
-      @text
-    end
-
-    def length
-      @text.length
-    end
+  class Text < String
   end
 end

data/lib/konjak/translation_unit.rb

@@ -30,6 +30,22 @@ module Konjak
       child_elements.count {|e| TranslationUnitVariant === e } >= 2
     end
 
+    def has_translation?(src_lang, target_lang)
+      src_lang?(src_lang) && has_variant_lang?(src_lang) && has_variant_lang?(target_lang)
+    end
+
+    def src_lang?(src_lang)
+      !self.src_lang || self.src_lang == '*all*' || self.src_lang == src_lang
+    end
+
+    def has_variant_lang?(lang)
+      variants.any? {|v| v.xml_lang == lang }
+    end
+
+    def variant(lang)
+      variants.detect {|v| v.xml_lang == lang }
+    end
+
     # FIXME
     #     Zero, one or more <note>, or <prop> elements in any order, followed by
     #     One or more <tuv> elements.

data/lib/konjak/translator/gtt_html_translate.rb

@@ -0,0 +1,47 @@
+module Konjak
+  class Translator
+    module GttHtmlTranslate
+      refine(Text) do
+        def gtt_tag_ns
+          scan(/\{(\d+)\}/).flatten.uniq
+        end
+
+        def compile_gtt_html_pattern
+          regexp = Regexp.escape(self)
+          gtt_tag_ns.each do |n|
+            regexp = regexp.sub(/\\\{#{n}\\\}/)    { "(?<n#{n}><(?<_#{n}>\\w+)[^>]*>)" }
+            regexp = regexp.gsub(/\\\{#{n}\\\}/)   { "\\k<n#{n}>" }
+            regexp = regexp.gsub(/\\\{\/#{n}\\\}/) { "</\\k<_#{n}>>" }
+          end
+          Regexp.compile(regexp)
+        end
+
+        def interpolate_gtt_html_pattern(match_data)
+          new_text = dup
+          gtt_tag_ns.each do |n|
+            new_text = new_text.gsub("{#{n}}", match_data["n#{n}"])
+            new_text = new_text.gsub("{/#{n}}", "</#{match_data["_#{n}"]}>")
+          end
+          new_text
+        end
+      end
+
+      refine(TranslationUnit) do
+        def translate(src_lang, target_lang, text)
+          pattern     = variant(src_lang).segment.text.compile_gtt_html_pattern
+          target_text = variant(target_lang).segment.text
+
+          texts = []
+          while true
+            head, match, tail = text.partition(pattern)
+            break if match.empty?
+            texts << head unless head.empty?
+            texts << TranslatedString.new(target_text.interpolate_gtt_html_pattern($~))
+            text = tail
+          end
+          texts << text
+        end
+      end
+    end
+  end
+end

data/lib/konjak/translator/text_translate.rb

@@ -0,0 +1,22 @@
+module Konjak
+  class Translator
+    module TextTranslate
+      refine(TranslationUnit) do
+        def translate(src_lang, target_lang, text)
+          s = variant(src_lang).segment.text
+          t = variant(target_lang).segment.text
+
+          texts = []
+          while true
+            head, match, tail = text.partition(s)
+            break if match.empty?
+            texts << head unless head.empty?
+            texts << TranslatedString.new(t)
+            text = tail
+          end
+          texts << text
+        end
+      end
+    end
+  end
+end

data/lib/konjak/translator/translated_string.rb

@@ -0,0 +1,6 @@
+module Konjak
+  class Translator
+    class TranslatedString < String
+    end
+  end
+end

data/lib/konjak/translator.rb

@@ -1,39 +1,34 @@
 require 'mem'
+require 'konjak/translator/gtt_html_translate'
+require 'konjak/translator/text_translate'
+require 'konjak/translator/translated_string'
 
 module Konjak
   class Translator
+
     include Mem
 
-    attr_reader :tmx, :src_lang, :target_lang
+    attr_reader :tmx, :src_lang, :target_lang, :options
 
-    def initialize(tmx, src_lang, target_lang)
-      @tmx = tmx
-      @src_lang = src_lang
+    def initialize(tmx, src_lang, target_lang, **options)
+      @tmx         = tmx
+      @src_lang    = src_lang
       @target_lang = target_lang
+      @options     = options
     end
 
     def translate(doc)
       translated_docs = [doc.dup]
       translation_units.each do |tu|
-        s = tu.variants.detect { |v| v.xml_lang == src_lang }.segment.text.to_s
-        t = tu.variants.detect { |v| v.xml_lang == target_lang }.segment.text.to_s
-        translated_docs.map! { |d|
-          next d if d.respond_to?(:translated)
-          next d if !d.include?(s)
-
-          ds = []
-          tail = nil
-          loop do
-            head, match, tail = d.partition(s)
-            ds << head
-            ds << t.dup.tap {|t| def t.translated; true; end }
-
-            break unless tail.include?(s)
-
-            d = tail
-          end
-          ds << tail
-          ds
+        translated_docs.map! { |text|
+          next text if text.is_a?(TranslatedString)
+
+          env = translate_env.dup
+          env.local_variable_set(:tu, tu)
+          env.local_variable_set(:src_lang, src_lang)
+          env.local_variable_set(:target_lang, target_lang)
+          env.local_variable_set(:text, text)
+          eval('tu.translate(src_lang, target_lang, text)', env)
         }.flatten!
       end
       translated_docs.join
@@ -41,13 +36,28 @@ module Konjak
 
     private
 
+    TRANSLATE_ENVS= {
+      text:     Class.new { using TextTranslate;    break binding },
+      gtt_html: Class.new { using GttHtmlTranslate; break binding }
+    }
+
+    def format
+      if TRANSLATE_ENVS.has_key?(options[:format])
+        options[:format]
+      else
+        :text
+      end
+    end
+
+    def translate_env
+      TRANSLATE_ENVS[format]
+    end
+
     def translation_units
       tmx.body.translation_units.select { |tu|
-        (!tu.src_lang || tu.src_lang == src_lang || tu.src_lang == '*all*') &&
-        tu.variants.any? {|v| v.xml_lang == src_lang } &&
-        tu.variants.any? {|v| v.xml_lang == target_lang }
+        tu.has_translation?(src_lang, target_lang)
       }.sort_by {|tu|
-        -tu.variants.detect { |v| v.xml_lang == src_lang }.segment.text.length
+        -tu.variant(src_lang).segment.text.length
       }
     end
     memoize :translation_units

data/lib/konjak/version.rb

@@ -1,3 +1,3 @@
 module Konjak
-  VERSION = "0.0.1"
+  VERSION = "0.0.2"
 end

data/lib/konjak.rb

@@ -36,13 +36,13 @@ require 'konjak/translator'
 
 module Konjak
   class << self
-    def parse(xml, **kw)
-      Parser.new.parse(xml, **kw)
+    def parse(xml, **options)
+      Parser.new.parse(xml, **options)
     end
 
-    def translate(doc, xml_or_tmx, src_lang, target_lang)
+    def translate(doc, xml_or_tmx, src_lang, target_lang, **options)
       tmx = xml_or_tmx.kind_of?(Tmx) ? xml_or_tmx : parse(xml_or_tmx)
-      Translator.new(tmx, src_lang, target_lang).translate(doc)
+      Translator.new(tmx, src_lang, target_lang, **options).translate(doc)
     end
   end
 end

data/spec/konjak_parse_spec.rb

@@ -162,7 +162,7 @@ describe Konjak do
   describe 'gtt' do
     let(:xml) { File.read('spec/fixtures/gtt.tmx') }
 
-    subject { tmx.body.translation_units.detect {|tu| tu.variants.detect {|v| v.segment.text.to_s == "\n\n& it's also example." } } }
+    subject { tmx.body.translation_units.detect {|tu| tu.variants.detect {|v| v.segment.text == "\n\n& it's also example." } } }
 
     context 'gtt: true' do
       let(:tmx) { Konjak.parse(xml, gtt: true) }

data/spec/konjak_translate_spec.rb

@@ -24,4 +24,21 @@ this is données (avec un caractère non standard: ).
 EXPECT
   end
 
+  context 'when format is GTT' do
+    let(:gtt_tmx) { Konjak.parse(File.read('spec/fixtures/gtt.tmx')) }
+
+    let(:doc) { <<GTT_HTML }
+This is <a href="http://example.com">example</a>.
+And This is <b>example</b>. Yey.
+And This is example.
+GTT_HTML
+
+    subject { Konjak.translate(doc, gtt_tmx, 'en', 'ja', format: :gtt_html) }
+
+    it { is_expected.to eq <<EXPECT }
+これは、 <a href="http://example.com">例</a> 。
+And これは、 <b>例</b> 。 Yey.
+And This is example.
+EXPECT
+  end
 end

data/spec/spec_helper.rb

@@ -1,3 +1,8 @@
+if ENV['CODECLIMATE_REPO_TOKEN']
+  require "codeclimate-test-reporter"
+  CodeClimate::TestReporter.start
+end
+
 require 'konjak'
 
 # spec/fixtures/sample.tmx from http://www.ttt.org/oscarstandards/tmx/#AppSample

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: konjak
 version: !ruby/object:Gem::Version
-  version: 0.0.1
+  version: 0.0.2
 platform: ruby
 authors:
 - Seiei Higa
@@ -162,6 +162,9 @@ files:
 - lib/konjak/translation_unit.rb
 - lib/konjak/translation_unit_variant.rb
 - lib/konjak/translator.rb
+- lib/konjak/translator/gtt_html_translate.rb
+- lib/konjak/translator/text_translate.rb
+- lib/konjak/translator/translated_string.rb
 - lib/konjak/unknown_tag.rb
 - lib/konjak/user_defined_encoding.rb
 - lib/konjak/version.rb