ckip_client 0.0.6 → 0.0.7

data/lib/CKIP_Client.rb

@@ -8,21 +8,14 @@ module CKIP
   class Client
     
     def self.get( sys , text )
-      begin
-        text.encode("Big5-UAO")
-      rescue Encoding::UndefinedConversionError
-        trans_text = text.encode("Big5-UAO", :undef => :replace).encode("UTF-8")
-        un_w = text.delete( trans_text )
-        puts "!!contains unsupported character: #{un_w}!!"
-        raise $!
-      end
       text_encoding = text.encoding.to_s
       unless ['Big5','Big5-UAO','UTF-8'].include? text_encoding
         raise 'Encoding ERROR!! CKIP_Client only support UTF-8 or Big5 or Big5-UAO encodings.'
       end
       input_encoding = (text_encoding == 'Big5-UAO')? 'Big5' : text_encoding
-      sst = 1.6 - 1020.0 / (text.size + 640)
+      sst = 2.0 - 2304.0 / (text.size + 1280)
       config = YAML::load( File.open( File.dirname(__FILE__) + "/config/#{sys}.yml" ).read )
+      sleep rand * 0.25 + 0.1
       request = "<?xml version=\"1.0\" ?>
 <wordsegmentation version=\"0.1\" charsetcode=\"#{input_encoding.downcase}\">
 <option showcategory=\"1\" />
@@ -30,29 +23,40 @@ module CKIP
 <text>#{text}</text>
 </wordsegmentation>"
       begin
-        Timeout::timeout(5) {
-          socket = TCPSocket.open( config['host'] , config['port'] )
-          socket.write( request )
-          @xml_result = socket.gets.force_encoding( text_encoding )
-          socket.close
+        time0 = Time.now
+        xml_result = Timeout::timeout(8.0 * (sst + 1.0)){
+          @socket = TCPSocket.open( config['host'] , config['port'] )
+          @socket.write( request )
+          @socket.gets.force_encoding( text_encoding )
         }
-        sleep (rand + 0.25) * sst
-        return @xml_result
+        time1 = (Time.now - time0)
+        sleep (rand + 0.5) * sst + time1 * 0.35
+        
+        if xml_result.valid_encoding?
+          return xml_result.encode!('UTF-8')
+        else
+          trans_text = xml_result.encode("UTF-32", :undef => :replace, :invalid => :replace).encode( text_encoding )
+          text2 = text.gsub(/[^[:word:]]+/ , "")
+          trans_text.each_char{ |c| text2.delete!(c) }
+          puts "!!contains unsupported character: #{text2}!!"
+          raise Encoding::InvalidByteSequenceError
+        end
       rescue Timeout::Error
-        puts "CKIP Connection Timeout!!!"
-        raise Timeout::Error
+        time1 = (Time.now - time0)
+        puts "!!!Timeout: waited for #{time1.round(2)}s and no response from CKIP server!!!"
+        raise $!
+      ensure
+        @socket.close
       end
     end
     
     def self.xml2str( xml )
-      xml.encode!('UTF-8')
-      sleep (rand * 0.1 + 0.05)
       if /<result>(.*?)<\/result>/m.match( xml )
         return $1.gsub(/<\/sentence>\r?\n?\t*?\s*?<sentence>/,"\n").gsub("\n　\n","\n\n").sub(/\t*?\s*?<sentence>　?/,'').sub(/<\/sentence>/,'').gsub("\n　", "\n")
       elsif /<processstatus code="\d">(.*?)<\/processstatus>/.match( xml )
         raise $1
       else
-        raise "XML result format error!!"
+        raise "XML return error!!"
       end
     end
     
@@ -61,7 +65,7 @@ module CKIP
   def self.segment( text , mode = nil )
     output = Client.xml2str( Client.get( 'segment' , text ) )
     if ['compact','neat'].include?( mode )
-      return output.gsub!(/\([A-Za-z]+\)/,'')
+      return output.gsub!(/\([A-Za-z_]+\)/,'')
     else
       return output
     end
@@ -69,10 +73,9 @@ module CKIP
   
   def self.parser( text , mode = nil )
     text.encode!('Big5-UAO') if text.encoding.to_s == 'UTF-8'
-    
     output = Client.xml2str( Client.get( 'parser' , text ) )
     if ['compact','neat'].include?( mode )
-      return output.gsub(/[A-Za-z]+?:/,'').gsub(/[A-Za-z]+?\(/,'(').gsub(/[A-Za-z]+?‧.+?\(/,'(').gsub!(/[#%]/,'').gsub!(/^\d+:\d+.\[\d+\]\s/,'').gsub(/\([A-Z]+?\)$/,'')
+      return output.gsub(/[A-Za-z_]+?:/,'').gsub(/[A-Za-z_]+?\(/,'(').gsub(/[A-Za-z_]+?‧.+?\(/,'(').gsub!(/[#%]/,'').gsub!(/^\d+:\d+.\[\d+\]\s/,'').gsub(/\([A-Z]+?\)$/,'')
     else
       return output
     end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: ckip_client
 version: !ruby/object:Gem::Version
-  version: 0.0.6
+  version: 0.0.7
   prerelease: 
 platform: ruby
 authors:
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2013-05-02 00:00:00.000000000 Z
+date: 2013-05-06 00:00:00.000000000 Z
 dependencies: []
 description: CKIP_CLient是連接中央研究院詞庫小組研發之中文斷詞系統與中文剖析系統的Ruby程式界面。感謝中央研究院詞庫小組多年來之研究成果！
 email: xxxooo.tw@gmail.com