sanzang 1.0.4 → 1.0.5

data/bin/sanzang

@@ -1,7 +1,7 @@
 #!/usr/bin/env ruby
-# -*- encoding: UTF-8 -*-
+# coding: UTF-8
 #--
-# Copyright (C) 2012 Lapis Lazuli Texts
+# Copyright (C) 2012-2013 Lapis Lazuli Texts
 #
 # This program is free software: you can redistribute it and/or modify it under
 # the terms of the GNU General Public License as published by the Free Software

data/lib/sanzang.rb

@@ -1,5 +1,4 @@
-#!/usr/bin/env ruby -w
-# -*- encoding: UTF-8 -*-
+# coding: UTF-8
 #--
 # Copyright (C) 2012-2013 Lapis Lazuli Texts
 #

data/lib/sanzang/batch_translator.rb

@@ -1,5 +1,4 @@
-#!/usr/bin/env ruby
-# -*- encoding: UTF-8 -*-
+# coding: UTF-8
 #--
 # Copyright (C) 2012-2013 Lapis Lazuli Texts
 #
@@ -36,12 +35,15 @@ module Sanzang
     # return value is an array containing all the output file paths.
     #
     def translate_batch(fpath_pairs, verbose = true, jobs = nil)
-      if not Sanzang::Platform.unix_processes?
-        jobs = 0
-      elsif not jobs
-        jobs = Sanzang::Platform.processor_count
+      options = {}
+      if RUBY_PLATFORM =~ /java/
+        options[:in_threads] = jobs || Sanzang::Platform.processor_count
+      elsif Sanzang::Platform.unix_processes?
+        options[:in_processes] = jobs || Sanzang::Platform.processor_count
+      else
+        options[:in_processes] = 0
       end
-      Parallel.map(fpath_pairs, :in_processes => jobs) do |f1,f2|
+      Parallel.map(fpath_pairs, options) do |f1,f2|
         translate_io(f1, f2)
         if verbose
           $stderr.write "[#{Process.pid}] #{File.expand_path(f2)} \n"

data/lib/sanzang/command/batch.rb

@@ -1,5 +1,4 @@
-#!/usr/bin/env ruby
-# -*- encoding: UTF-8 -*-
+# coding: UTF-8
 #--
 # Copyright (C) 2012-2013 Lapis Lazuli Texts
 #

data/lib/sanzang/command/reflow.rb

@@ -1,5 +1,4 @@
-#!/usr/bin/env ruby
-# -*- encoding: UTF-8 -*-
+# coding: UTF-8
 #--
 # Copyright (C) 2012-2013 Lapis Lazuli Texts
 #

data/lib/sanzang/command/sanzang_cmd.rb

@@ -1,5 +1,4 @@
-#!/usr/bin/env ruby
-# -*- encoding: UTF-8 -*-
+# coding: UTF-8
 #--
 # Copyright (C) 2012-2013 Lapis Lazuli Texts
 #

data/lib/sanzang/command/translate.rb

@@ -1,5 +1,4 @@
-#!/usr/bin/env ruby
-# -*- encoding: UTF-8 -*-
+# coding: UTF-8
 #--
 # Copyright (C) 2012-2013 Lapis Lazuli Texts
 #

data/lib/sanzang/platform.rb

@@ -1,5 +1,4 @@
-#!/usr/bin/env ruby
-# -*- encoding: UTF-8 -*-
+# coding: UTF-8
 #--
 # Copyright (C) 2012-2013 Lapis Lazuli Texts
 #

data/lib/sanzang/text_formatter.rb

@@ -1,5 +1,4 @@
-#!/usr/bin/env ruby
-# -*- encoding: UTF-8 -*-
+# coding: UTF-8
 #--
 # Copyright (C) 2012-2013 Lapis Lazuli Texts
 #

data/lib/sanzang/translation_table.rb

@@ -1,5 +1,4 @@
-#!/usr/bin/env ruby
-# -*- encoding: UTF-8 -*-
+# coding: UTF-8
 #--
 # Copyright (C) 2012-2013 Lapis Lazuli Texts
 #
@@ -29,7 +28,7 @@ module Sanzang
     # summarized as follows:
     #
     # - Each line of text is a record for a translation rule.
-    # - Each record begins with "~|" and ends with "|~".
+    # - Each record may begin with "~|" and end with "|~".
     # - Fields in the record are separated by the "|" character.
     # - The first field contains the term in the source language.
     # - Subsequent fields are equivalent terms in destination languages.
@@ -43,15 +42,11 @@ module Sanzang
     def initialize(rules)
       contents = rules.kind_of?(String) ? rules : rules.read
       contents.encode!(Encoding::UTF_8)
-      @encoding = contents.encoding
-
-      left = "~|"
-      right = "|~"
-      separator = "|"
-
-      @records = contents.gsub("\r", "").split("\n").collect do |rec|
-        rec.strip.gsub(left, "").gsub(right, "").split(separator)
-      end
+      contents.strip!
+      contents.gsub!(/^\s*|\s*$|\r/, "")
+      contents.gsub!("~|", "")
+      contents.gsub!("|~", "")
+      @records = contents.split("\n").collect {|r| r.split("|") }
 
       if @records.length < 1
         raise "Table must have at least 1 row"
@@ -59,9 +54,9 @@ module Sanzang
         raise "Table must have at least 2 columns"
       end
 
-      @width = records[0].length
-      0.upto(@records.length - 1) do |i|
-        if @records[i].length != @width
+      width = records[0].length
+      @records.each do |r|
+        if r.length != width
           raise "Column mismatch: Line #{i + 1}"
         end
       end
@@ -75,6 +70,12 @@ module Sanzang
       @records[index]
     end
 
+    # The text encoding used for all translation table data
+    #
+    def encoding
+      Encoding::UTF_8
+    end
+
     # Find a record by the source language term (first column).
     #
     def find(term)
@@ -97,9 +98,5 @@ module Sanzang
     #
     attr_reader :records
 
-    # The text encoding used for all translation table data
-    #
-    attr_reader :encoding
-
   end
 end

data/lib/sanzang/translator.rb

@@ -1,5 +1,4 @@
-#!/usr/bin/env ruby
-# -*- encoding: UTF-8 -*-
+# coding: UTF-8
 #--
 # Copyright (C) 2012-2013 Lapis Lazuli Texts
 #

data/lib/sanzang/version.rb

@@ -1,5 +1,4 @@
-#!/usr/bin/env ruby
-# -*- encoding: UTF-8 -*-
+# coding: UTF-8
 #--
 # Copyright (C) 2012-2013 Lapis Lazuli Texts
 #
@@ -20,6 +19,6 @@ module Sanzang
 
   # Current version number of Sanzang
   #
-  VERSION = "1.0.4"
+  VERSION = "1.0.5"
 
 end

data/test/tc_reflow_encodings.rb

@@ -1,5 +1,4 @@
-#!/usr/bin/env ruby
-# -*- encoding: UTF-8 -*-
+# coding: UTF-8
 
 require "test/unit"
 

data/test/tc_simple_translation.rb

@@ -1,5 +1,4 @@
-#!/usr/bin/env ruby
-# -*- encoding: UTF-8 -*-
+# coding: UTF-8
 
 require "test/unit"
 

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: sanzang
 version: !ruby/object:Gem::Version
-  version: 1.0.4
+  version: 1.0.5
   prerelease: 
 platform: ruby
 authors:
@@ -9,32 +9,31 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2013-07-25 00:00:00.000000000 Z
+date: 2013-10-13 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: parallel
   requirement: !ruby/object:Gem::Requirement
     none: false
     requirements:
-    - - ! '>='
+    - - ~>
       - !ruby/object:Gem::Version
-        version: 0.5.19
+        version: '0.8'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     none: false
     requirements:
-    - - ! '>='
+    - - ~>
       - !ruby/object:Gem::Version
-        version: 0.5.19
+        version: '0.8'
 description: Sanzang is a program built for machine translation of natural languages.
   This application is particularly suitable as a translation aid for CJK languages
   including ancient texts. The translation method is rule-based, and translation rules
   are stored in flat files as delimited text. This program can also utilize multiprocessing
   to naturally scale to multiple processors and processor cores. Sanzang is available
   under the GNU GPL, version 3.
-email:
-- lapislazulitexts@gmail.com
+email: lapislazulitexts@gmail.com
 executables:
 - sanzang
 extensions: []
@@ -90,12 +89,13 @@ required_rubygems_version: !ruby/object:Gem::Requirement
   - - ! '>='
     - !ruby/object:Gem::Version
       version: '0'
-requirements: []
+requirements:
+- parallel ~> 0.8
 rubyforge_project: 
 rubygems_version: 1.8.23
 signing_key: 
 specification_version: 3
-summary: Simple rule-based machine translation system.
+summary: Machine translation for CJK languages
 test_files:
 - test/tc_reflow_encodings.rb
 - test/tc_simple_translation.rb