konjak 0.0.7 → 0.0.8

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: c2fc4ee1a4ae07170e1d4d0af6fbed84b6ab60e5
-  data.tar.gz: 93ca5aa335e0b012a9f9f727926ed4918099fa5e
+  metadata.gz: cdc98142b5ed2a6814a96cf76c8d9c0f7bd2c5d2
+  data.tar.gz: f286772061e56d381e8f4f227e47ae9bfa259362
 SHA512:
-  metadata.gz: 9ebaa19e5a5598a4053de05f718bf643776c2611bf2d4006ddae8195806e593f8f35b43471f943e1d46ee8f4dea4f617ccc607f621f4ec5351c30f4110725fe2
-  data.tar.gz: 30be6bc899940236afe204122fcc916e4441eb2125db0e31a2d0698ec70f2b1fe5aad70a8fa226caa348faf410de2bcef5ea6a26e37f8232233a978c6cc4a0c2
+  metadata.gz: ca86d1d1693776a7b3a5e57c687daf199d879cbb2e01c6ad6d806def612d1cc01e1c0938b91cf87b837406cb4bdbeee713a4c5d2043c1319faf3e22fc6798beb
+  data.tar.gz: 4339756d7df8a10a218d906cb12b85af05389056ebddef23219f9f6ecdd75f6c12f9010a4b5164faed85ee72a7e2a3f80f44151dde99cb51b7c8438ce3b92888

data/lib/konjak/segment/gtt.rb

@@ -10,6 +10,7 @@ module Konjak
           regexp = regexp.gsub(/\\\{#{n}\\\}/)   { "\\k<n#{n}>" }
           regexp = regexp.gsub(/\\\{\/#{n}\\\}/) { "</\\k<_#{n}>>" }
         end
+        regexp = regexp.gsub(/(\\\s|\n)/m)     { '\s+' }
         Regexp.compile(regexp)
       end
 

data/lib/konjak/segment.rb

@@ -15,6 +15,12 @@ module Konjak
       [Text, BeginPairedTag, EndPairedTag, IsolatedTag, Placeholder, Highlight].any? {|c| c === element }
     end
 
+    def compile_pattern
+      regexp = Regexp.escape(text)
+      regexp = regexp.gsub(/(\\\s|\n)/m) { '\s+' }
+      Regexp.compile(regexp)
+    end
+
     def translation_unit
       TranslationUnit.new(translation_unit_variant.parent)
     end

data/lib/konjak/tmx_segmentor/text_strategy.rb

@@ -9,11 +9,11 @@ module Konjak
 
       def split(translation_unit, text)
         segment = translation_unit.variant(@lang).segment
-        segment_text = segment.text
+        pattern = segment.compile_pattern
 
         texts = []
         while true
-          head, match, tail = text.partition(segment_text)
+          head, match, tail = text.partition(pattern)
           break if match.empty? || text.length < min_segment_length
           texts << head unless head.empty?
 

data/lib/konjak/version.rb

@@ -1,3 +1,3 @@
 module Konjak
-  VERSION = "0.0.7"
+  VERSION = "0.0.8"
 end

data/spec/konjak_translate_spec.rb

@@ -8,20 +8,30 @@ this is data (with a non-standard character: ).
 this is data (with a non-standard character: ).
 DOC
 
-  subject { Konjak.translate(doc, sample_tmx, 'EN', 'FR-CA').join }
-
-  it { is_expected.to eq <<EXPECT }
+  let(:expected) { <<EXPECT }
 this is données (avec un caractère non standard: ).
 this is données (avec un caractère non standard: ).
 EXPECT
 
+  subject { Konjak.translate(doc, sample_tmx, 'EN', 'FR-CA').join }
+
+  it { is_expected.to eq expected }
+
+  context 'when blanks between words is not exactly match' do
+    let(:doc) { <<DOC }
+this is data         (with a non-standard
+
+character: ).
+this is data (with a non-standard character: ).
+DOC
+
+    it { is_expected.to eq expected }
+  end
+
   context 'when passed Tmx' do
     subject { Konjak.translate(doc, Konjak.parse(sample_tmx), 'EN', 'FR-CA').join }
 
-    it { is_expected.to eq <<EXPECT }
-this is données (avec un caractère non standard: ).
-this is données (avec un caractère non standard: ).
-EXPECT
+    it { is_expected.to eq expected }
   end
 
   context 'when format is GTT' do
@@ -32,13 +42,29 @@ 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).join }
-
-    it { is_expected.to eq <<EXPECT }
+    let(:expected) { <<EXPECT }
 これは、 <a href="http://example.com">例</a> 。
 And これは、 <b>例</b> 。 Yey.
 And This is example.
 EXPECT
+
+    subject { Konjak.translate(doc, gtt_tmx, 'en', 'ja', format: :gtt_html).join }
+
+    it { is_expected.to eq expected }
+
+    context 'when blanks between words is not exactly match' do
+      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
+
+      it { is_expected.to eq expected }
+    end
+
   end
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: konjak
 version: !ruby/object:Gem::Version
-  version: 0.0.7
+  version: 0.0.8
 platform: ruby
 authors:
 - Seiei Higa
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2015-06-22 00:00:00.000000000 Z
+date: 2015-06-23 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: mem