konjac 0.1.3 → 0.1.4

data/.gitignore

@@ -2,4 +2,5 @@
 .bundle
 .DS_Store
 Gemfile.lock
+doc/
 pkg/*

data/README.md

@@ -63,13 +63,15 @@ Use multiple dictionaries:
 
     konjac translate financial_report_en.txt into japanese using finance
 
-Extract text from a DOCX document (creates a plain-text `test.tags` file):
+Extract text from a DOCX document (creates a plain-text `test.konjac` file from
+`test.docx`):
 
-    konjac extract test.docx
+    konjac extract test
 
-Import tags file back into DOCX document:
+Import tags file back into DOCX document (file created is named
+`test_imported.docx`):
 
-    konjac import test.docx
+    konjac import test
 
 Extended Example
 ----------------
@@ -108,7 +110,7 @@ Extended Example
         en: ""
         ja: !ruby/regexp /\s(?=[.,:]|$)/
 
-~/.konjac/test_en.txt
+~/.konjac/test_en.txt:
 
     I like dogs.
 

data/lib/konjac/cli/sub_command.rb

@@ -0,0 +1,20 @@
+module Konjac
+  module CLI
+    class SubCommand
+      attr :name, :aliases, :description
+
+      def initialize(name, aliases, description, &block)
+        @name, @aliases, @description = name, aliases, description
+        @block = block
+      end
+
+      def alias_text
+        @aliases.join ", "
+      end
+
+      def execute
+        @block.call
+      end
+    end
+  end
+end

data/lib/konjac/cli/sub_command_manager.rb

@@ -0,0 +1,77 @@
+module Konjac
+  module CLI
+    class SubCommandManager
+      def initialize
+        @sub_commands ||= []
+      end
+
+      def add(name, aliases, description, &block)
+        @sub_commands << SubCommand.new(name, aliases, description, &block)
+      end
+
+      def execute(sub_command)
+        sub_command = sub_command.to_sym
+
+        @sub_commands.each do |sc|
+          if sc.name == sub_command || sc.aliases.include?(sub_command)
+            sc.execute
+            return sc.name
+          end
+        end
+
+        return nil
+      end
+
+      def all
+        @sub_commands
+      end
+
+      def get_lengths
+        @lengths = { :name => 8, :aliases => 7, :description => 11 }
+
+        @sub_commands.each do |sc|
+          if sc.name.length > @lengths[:name]
+            @lengths[:name] = sc.name.length
+          end
+          if sc.alias_text.length > @lengths[:aliases]
+            @lengths[:aliases] = sc.alias_text.length
+          end
+          if sc.description.length > @lengths[:description]
+            @lengths[:description] = sc.description.length
+          end
+        end
+      end
+
+      def to_s
+        get_lengths
+        text = "\n"
+
+        header_lines = "%s  %s  %s\n" % [
+          "-" * @lengths[:name],
+          "-" * @lengths[:aliases],
+          "-" * @lengths[:description]
+        ]
+
+        # Create headers
+        text << header_lines
+        text << "%s  %s  %s\n" % [
+          "Commands".ljust(@lengths[:name]),
+          "Aliases".ljust(@lengths[:aliases]),
+          "Description".ljust(@lengths[:description])
+        ]
+        text << header_lines
+
+        # Write descriptions of each subcommand
+        @sub_commands.each do |sc|
+          text << "%s  %s  %s\n" % [
+            sc.name.to_s.ljust(@lengths[:name]),
+            sc.alias_text.to_s.ljust(@lengths[:aliases]),
+            sc.description.to_s.ljust(@lengths[:description])
+          ]
+        end
+
+        text
+      end
+    end
+  end
+end

data/lib/konjac/cli.rb

@@ -1,28 +1,43 @@
 module Konjac
   module CLI
+    autoload :SubCommand,        "konjac/cli/sub_command"
+    autoload :SubCommandManager, "konjac/cli/sub_command_manager"
+
     class << self
+      def init_sub_commands
+        @valid_commands = SubCommandManager.new
+        @valid_commands.add :extract, [:export, :e, :x], "Extract text from a DOCX file" do
+          Word.extract_docx_tags(ARGV)
+        end
+        @valid_commands.add :help, [:h, :"-h", :"?"], "Show help" do
+          show_help
+        end
+        @valid_commands.add :import, [:i, :m], "Import text back into a DOCX file" do
+          Word.import_docx_tags(ARGV)
+        end
+        @valid_commands.add :translate, [:t], "Translate a file" do
+          translate
+        end
+      end
+
       def start
+        init_sub_commands
         show_help if ARGV.empty?
 
         # Subcommand
-        case ARGV.shift
-        when "translate"
-          translate
-        when "extract", "export"
-          Word.extract_docx_tags(ARGV)
-        when "import"
-          Word.import_docx_tags(ARGV)
-        when "add"
-        when "help"
+        sub_command = ARGV.shift
+        result = @valid_commands.execute(sub_command)
+
+        if result.nil?
           show_help
-        else
-          raise InvalidCommandError.new("Valid commands are translate or add")
+          raise InvalidCommandError.new("Invalid subcommand: #{sub_command}")
         end
+
+        result
       end
 
       def show_help
-        puts "Help"
-        exit 0
+        puts @valid_commands.to_s
       end
 
       private

data/lib/konjac/dictionary.rb

@@ -16,14 +16,13 @@ module Konjac
         from_lang = from_lang.to_s
         to_lang   = to_lang.to_s
 
+        # Ignore everything if we've been here before
+        return @pairs if loaded?(from_lang, to_lang, dictionaries) || opts[:force]
+        
         # Build a regex template for the from language
         from_template = build_regex_template(from_lang)
         to_template   = build_replacement_template(from_lang, to_lang)
 
-        # Get full path from name
-        
-        return @pairs if loaded?(from_lang, to_lang, dictionaries) || opts[:force]
-
         # Save variables to cache so we can avoid repetitive requests
         cache_load from_lang, to_lang, dictionaries
         
@@ -46,40 +45,51 @@ module Konjac
         # Build a list of search and replace pairs
         @pairs = []
         @dictionary.each do |term|
-          if term.has_key?(to_lang)
-            # Build to term depending on whether it's a hash for complex
-            # matches or a simple string
-            if term[to_lang].is_a?(Hash)
-              to_term = term[to_lang][to_lang]
-
-              if term[to_lang].has_key?(from_lang)
-                from_term = term[to_lang][from_lang]
-              end
-            else
-              to_term = term[to_lang]
-            end
+          pair = extract_pair_from_term(term, from_lang, to_lang, from_template, to_template)
+          @pairs << pair unless pair.nil?
+        end
 
-            # Build from term if it doesn't already exist
-            if term.has_key?(from_lang) && from_term.nil?
-              from_term = term[from_lang]
-            end
+        @pairs
+      end
 
-            unless from_term.nil?
-              # Build a regular expression if it isn't already one
-              # Note that this will apply word boundary rules, so to avoid them
-              # create a regular expression in the dictionary file
-              unless from_term.is_a?(Regexp)
-                from_term = Regexp.new(from_template % from_term)
-              end
+      def extract_pair_from_term(term, from_lang, to_lang, from_template, to_template)
+        if term.has_key?(to_lang)
+          # Build to term depending on whether it's a hash for complex
+          # matches or a simple string
+          if term[to_lang].is_a?(Hash)
+            to_term = term[to_lang][to_lang]
 
-              to_term = to_template % to_term unless to_term =~ BLANK
+            if term[to_lang].has_key?(from_lang)
+              from_term = term[to_lang][from_lang]
+            end
+          else
+            to_term = term[to_lang]
+          end
 
-              @pairs << [ from_term, to_term ]
+          # Build from term if it doesn't already exist
+          if term.has_key?(from_lang) && from_term.nil?
+            from_term = term[from_lang]
+          end
+
+          unless from_term.nil?
+            # Build a regular expression if it isn't already one.
+            # Note that this will apply word boundary rules, so to avoid them
+            # create a regular expression in the dictionary file.
+            # Matching is case-insensitive unless the expression contains a
+            # capital letter.
+            unless from_term.is_a?(Regexp)
+              from_term = Regexp.new(from_template % from_term,
+                                     ("i" unless from_term =~ /[A-Z]/))
             end
+
+            to_term = to_template % to_term unless to_term =~ BLANK
+
+            return [ from_term, to_term ]
           end
         end
 
-        @pairs
+        # Return nil if no term could be constructed
+        return nil
       end
 
       # Verifies whether the dictionary exists on the specified path, creating

data/lib/konjac/version.rb

@@ -1,3 +1,3 @@
 module Konjac
-  VERSION = "0.1.3"
+  VERSION = "0.1.4"
 end

data/spec/cli_spec.rb

@@ -3,8 +3,15 @@ require File.dirname(__FILE__) + "/spec_helper"
 require "tempfile"
 
 describe CLI do
+  # Super hacky way of setting a constant without setting off a warnings that the
+  # constant has already been initialized
+  def set_argv(*params)
+    Object.__send__ :remove_const, :ARGV
+    Object.const_set :ARGV, params
+  end
+
   it "should fail on an invalid subcommand" do
-    ARGV = ["invalid"]
+    set_argv "invalid"
     lambda { CLI.start }.should raise_error InvalidCommandError
   end
 
@@ -14,8 +21,8 @@ describe CLI do
       @dictionary = Tempfile.new(["dict", ".yml"])
       @dictionary.write <<-eos
                           -
-                           en: dogs
-                           ja: 犬
+                            en: dogs
+                            ja: 犬
                         eos
       @dictionary.rewind
 
@@ -23,14 +30,17 @@ describe CLI do
       @english = Tempfile.new(["english", "_en.txt"])
       @english.write "I like dogs."
       @english.rewind
+
+      # Set ARGV
+      set_argv "translate", @english.path, "into", "japanese", "using",
+        @dictionary.path
     end
 
     it "should correctly translate English text" do
       begin
-        ARGV = ["translate", @english.path, "into", "japanese", "using", @dictionary.path]
         CLI.start
         converted_path = Utils.build_converted_file_name(@english.path, "en", "ja")
-        puts File.read(@english.path).should == "I like dogs."
+        File.read(@english.path).should == "I like dogs."
         File.read(converted_path).should == "I like 犬.\n"
       ensure
         @dictionary.close!
@@ -38,5 +48,9 @@ describe CLI do
         File.delete converted_path
       end
     end
+
+    it "should return the name of the sub_command" do
+      CLI.start.should == :translate
+    end
   end
 end

data/spec/dictionary_spec.rb

@@ -7,14 +7,82 @@ describe Dictionary do
     @dictionary = Tempfile.new(["dict", ".yml"])
     @dictionary.write <<-eos
                         -
-                         en: dogs
-                         ja: 犬
+                          en: dogs
+                          ja: 犬
+                        -
+                          en: cat
+                          ja:
+                            ja: 猫
+                            en: cats?
+                        -
+                          ja:
+                            ja: 。
+                            en: !ruby/regexp /\\.\\s?/
+                        -
+                          en: mouth
+                          es: boca
+                        -
+                          es: mano
+                          ja: 手
+                        -
+                          en:
+                            en: book
+                            ja: 本
+                        -
+                          en: North Carolina
+                          ja: ノースカロライナ
                       eos
     @dictionary.rewind
   end
 
-  it "should correctly load YAML" do
-    Dictionary.load "en", "ja", [@dictionary.path]
-    Dictionary.pairs.should == [[/\bdogs\b/, "犬"]]
+  describe "when converting from English to Japanese" do
+    before :each do
+      Dictionary.load "en", "ja", [@dictionary.path]
+    end
+
+    it "should correctly load a simple term" do
+      Dictionary.pairs.first.should == [/\bdogs\b/i, "犬"]
+    end
+
+    it "should correctly load a slightly more complex regular expression" do
+      Dictionary.pairs[1].should == [/\bcats?\b/i, "猫"]
+    end
+
+    it "should correctly load a manual regular expression" do
+      Dictionary.pairs[2].should == [/\.\s?/, "。"]
+    end
+
+    it "should ignore examples for which the source language cannot be found" do
+      Dictionary.pairs.keep_if { |t| t[0] == /\bmano\b/i }.empty?.should == true
+    end
+
+    it "should ignore examples for which the target language cannot be found" do
+      Dictionary.pairs.keep_if { |t| t[1] == "book" }.empty?.should == true
+    end
+
+    it "should only have a case-insensitive search if the search contains capital letters" do
+      Dictionary.pairs[0][0].should == /\bdogs\b/i
+      Dictionary.pairs[3][0].should == /\bNorth Carolina\b/
+    end
+  end
+
+  describe "when converting from Japanese to English" do
+    before :each do
+      Dictionary.load "ja", "en", [@dictionary.path]
+    end
+
+    it "should add whitespace to term replacement" do
+      Dictionary.pairs.first.should == [/犬/i, "dogs "]
+    end
+  end
+
+  describe "when converting from English to Spanish" do
+    before :each do
+      Dictionary.load "en", "es", [@dictionary.path]
+    end
+
+    it "should not add whitespace to term replacement" do
+      Dictionary.pairs[0].should == [/\bmouth\b/i, "boca"]
+    end
   end
 end

data/spec/language_spec.rb

@@ -11,6 +11,10 @@ describe Language do
     Language.find(:jpn).should == :ja
   end
 
+  it "should fail to find a made-up language" do
+    lambda { Language.find(:black_speech) }.should raise_error InvalidLanguageError
+  end
+
   it "should return equal results for both strings and symbols" do
     Language.find(:english).should == Language.find("english")
   end

data/spec/tag_spec.rb

@@ -6,15 +6,18 @@ describe Tag do
   before :each do
     @tags_file = Tempfile.new(["tags", ".tags"])
     @tags_file.write <<-eos.gsub(/^\s+/, "")
-                       [[KJ-1]]
+                       [[KJ-0]]
                        > 犬
                        dog
-                       [[KJ-2]]
+                       [[KJ-1]]
                        > 何ですか。
                        What is it?
-                       [[KJ-3]]
+                       [[KJ-2]]
                        > 空白
-                       [[KJ-6]]
+                       [[KJ-3]]  #=> comment
+                       > コメント
+                       Comment
+                       [[KJ-5]]
                        > 以上
                        -- end --
                      eos
@@ -22,26 +25,37 @@ describe Tag do
     @manager = TagManager.new(@tags_file.path)
   end
 
-  it "should accurately read a tag file" do
+  it "should accurately read a tag" do
     @manager.all.should_not == nil
-    @manager[0].index.should == 1
+    @manager[0].index.should == 0
     @manager[0].original.should == "犬"
     @manager[0].translated.should == "dog"
   end
 
-  it "should succeed reading multiple lines" do
-    @manager[1].index.should == 2
+  it "should succeed reading multiple tags" do
+    @manager[1].index.should == 1
     @manager[1].original.should == "何ですか。"
     @manager[1].translated.should == "What is it?"
   end
 
   it "should ignore blank translations" do
-    @manager[2].index.should == 3
+    @manager[2].index.should == 2
     @manager[2].original.should == "空白"
     @manager[2].translated.should == nil
   end
 
+  it "should know whether a tag has been translated" do
+    @manager[0].translated?.should == true
+    @manager[2].translated?.should == false
+  end
+
+  it "should ignore comments and additional info after an index tag" do
+    @manager[3].index.should == 3
+    @manager[3].original.should == "コメント"
+    @manager[3].translated.should == "Comment"
+  end
+
   it "should skip over blank indexes" do
-    @manager[3].index.should == 6
+    @manager[4].index.should == 5
   end
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: konjac
 version: !ruby/object:Gem::Version
-  version: 0.1.3
+  version: 0.1.4
   prerelease: 
 platform: ruby
 authors:
@@ -13,7 +13,7 @@ date: 2012-01-13 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: nokogiri
-  requirement: &70101148958680 !ruby/object:Gem::Requirement
+  requirement: &70265140795800 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -21,10 +21,10 @@ dependencies:
         version: '0'
   type: :runtime
   prerelease: false
-  version_requirements: *70101148958680
+  version_requirements: *70265140795800
 - !ruby/object:Gem::Dependency
   name: bundler
-  requirement: &70101148958040 !ruby/object:Gem::Requirement
+  requirement: &70265140794440 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -32,10 +32,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *70101148958040
+  version_requirements: *70265140794440
 - !ruby/object:Gem::Dependency
   name: rspec
-  requirement: &70101148683020 !ruby/object:Gem::Requirement
+  requirement: &70265140793660 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -43,10 +43,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *70101148683020
+  version_requirements: *70265140793660
 - !ruby/object:Gem::Dependency
   name: sdoc
-  requirement: &70101148682400 !ruby/object:Gem::Requirement
+  requirement: &70265140792780 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -54,10 +54,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *70101148682400
+  version_requirements: *70265140792780
 - !ruby/object:Gem::Dependency
   name: autotest
-  requirement: &70101148681900 !ruby/object:Gem::Requirement
+  requirement: &70265140791700 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -65,10 +65,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *70101148681900
+  version_requirements: *70265140791700
 - !ruby/object:Gem::Dependency
   name: autotest-fsevent
-  requirement: &70101148681460 !ruby/object:Gem::Requirement
+  requirement: &70265140790860 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -76,10 +76,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *70101148681460
+  version_requirements: *70265140790860
 - !ruby/object:Gem::Dependency
   name: autotest-growl
-  requirement: &70101148681040 !ruby/object:Gem::Requirement
+  requirement: &70265140789760 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -87,7 +87,7 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *70101148681040
+  version_requirements: *70265140789760
 description: A Ruby command-line utility for translating files using a YAML wordlist
 email:
 - bryan.mckelvey@gmail.com
@@ -107,6 +107,8 @@ files:
 - konjac.gemspec
 - lib/konjac.rb
 - lib/konjac/cli.rb
+- lib/konjac/cli/sub_command.rb
+- lib/konjac/cli/sub_command_manager.rb
 - lib/konjac/dictionary.rb
 - lib/konjac/exception.rb
 - lib/konjac/language.rb