konjak 0.0.17 → 0.0.18

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 5fbbfd98e759c97649e0296aeeed743df368383a
-  data.tar.gz: 45a1b8046c4a4db79712623cf97c57a57f050205
+  metadata.gz: c4ffc2d0548951d62d07f5612bd4399e8a8f6519
+  data.tar.gz: 77a83feb0b806e2e6a87b56246c7062fa175f364
 SHA512:
-  metadata.gz: 8831f0b9af534fb01182de0d11d3e96a62c62ce255dfabc843ed3e3452f3052a7fbb0e22b4937275889fa8fefaf75b2f63ee4623e1548a60d005a9a33cfd9c5f
-  data.tar.gz: 1a82b13769650fba60bfdbc08450cd5855923714e3025552e22fb9f7f0ef48add62251b797c9edca13ead13b79abd1823077c2f544aa7324b39e064e57568c33
+  metadata.gz: c92f4f5d4328b6fe29e5a92451b7e648796c37ec7c044a58ecfbb70e7620cf67494fce6a8052bbce6901caa85a529a5d35001ccafe236c7dd0b97efcd7e399d3
+  data.tar.gz: 995318581a91f8fbc41ddf97258777ba3942e6e7c69ab1b303208be8d7f1aecca44c2b0561c1a10adc70f5055e583223f97ad8fcca7787a8165f85ff37f8a8e8

data/lib/konjak/tmx_segmentor/strategy.rb

@@ -17,7 +17,7 @@ module Konjak
 
       def segmentize(text)
         segments = [text]
-        translation_units.each do |translation_unit|
+        translation_units(text).each do |translation_unit|
           segment = translation_unit.variant(@lang).segment
 
           pat = compile_pattern(segment)
@@ -35,13 +35,20 @@ module Konjak
       private
 
       def default_options
-        {min_segment_length: 10}
+        {
+          min_segment_length: 10,
+          max_segment_length: nil
+        }
       end
 
       def min_segment_length
         @options[:min_segment_length]
       end
 
+      def max_segment_length
+        @options[:max_segment_length]
+      end
+
       def split(pat, segment, text)
         texts = []
         while true
@@ -62,26 +69,53 @@ module Konjak
         texts << text
       end
 
-      def translation_units
-        @tmx.body.translation_units.sort_by {|tu|
+      def translation_units(text)
+        tus = @tmx.body.translation_units
+
+        tus.select! {|tu|
+          segment = tu.variant(@lang).segment
+          segment_length = segment.text.length
+
+          next false if segment_length < min_segment_length
+          next false if max_segment_length && max_segment_length < segment_length
+
+          text =~ compile_pattern(tu.variant(@lang).segment)
+        }
+
+        simular_translation_units_map = {}
+
+        tus.sort_by! {|tu|
+          tu_segment = tu.variant(@lang).segment
+          segment_text = tu_segment.text
+
+          unless simular_translation_units_map[segment_text]
+            simular_translation_units = tus.select {|tu2|
+              tu2.variant(@lang).segment.text.include?(segment_text)
+            }.sort_by! {|tu2| tu2.variant(@lang).segment.text.size }
+
+            simular_translation_units.each do |tu2|
+              simular_translation_units_map[tu2.variant(@lang).segment.text] = simular_translation_units
+            end
+          end
+
+          rank = simular_translation_units_map[segment_text].index {|tu2|
+            tu2.variant(@lang).segment.text == segment_text
+          }
+
           # GTTの場合
           translation_timestamp = nil
-
           if tm_entry = tu.at('entry_metadata').try(:at, 'tm_entry')
             source_info = tm_entry.at('source_info')
-            if source_info.try(:at, 'source_lang').try(:text) == @lang && source_info.try(:at, 'source').try(:text) == tu.variant(@lang).segment.text
+            if source_info.try(:at, 'source_lang').try(:text) == @lang && source_info.try(:at, 'source').try(:text) == segment_text
               translation_timestamp = tm_entry.at('translation').try(:attr, 'translation_timestamp').to_i
             end
           end
-
           translation_timestamp ||= 0
 
-          segment_length = tu.variant(@lang).segment.text.length
-
-          [-translation_timestamp, -segment_length]
-        }.reject {|tu|
-          tu.variant(@lang).segment.text.length < min_segment_length
+          [-rank, -translation_timestamp, -segment_text.length]
         }
+
+        tus
       end
     end
   end

data/lib/konjak/version.rb

@@ -1,3 +1,3 @@
 module Konjak
-  VERSION = "0.0.17"
+  VERSION = "0.0.18"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: konjak
 version: !ruby/object:Gem::Version
-  version: 0.0.17
+  version: 0.0.18
 platform: ruby
 authors:
 - Seiei Higa
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2015-09-17 00:00:00.000000000 Z
+date: 2015-09-18 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: activesupport
@@ -216,7 +216,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
 requirements: []
 rubyforge_project: 
-rubygems_version: 2.4.5
+rubygems_version: 2.4.5.1
 signing_key: 
 specification_version: 4
 summary: TMX(Translation Memory exChange) tools for ruby