any2tmx 1.0.0

checksums.yaml

@@ -0,0 +1,7 @@
+---
+SHA1:
+  metadata.gz: e34e73e877005effaf94729c0d3929359ee04e96
+  data.tar.gz: 5778d82075a5db914398b25af2c19e69885fd043
+SHA512:
+  metadata.gz: 7e7539b3f90f17c4b94e22838ad0d9325ed2a0201a03aba27dfff39f15422ca8c4d24d9b70130e0a702a1748b4a35a500daa3b7f1b35a60fa17fe11c18538926
+  data.tar.gz: 6658db3c6acd69e5fe0cfbb1013becffee73e159cee8b98f7483adca03448543dfdfea8f18a609af1754ac07e28db7fe4c1da12bd52c771abd178bacc05f87e4

data/Gemfile

@@ -0,0 +1,8 @@
+source 'http://rubygems.org'
+
+gemspec
+
+group :development, :test do
+  gem 'rake'
+  gem 'pry-nav'
+end

data/History.txt

@@ -0,0 +1,8 @@
+== 1.0.0
+
+* Birthday!
+
+== 1.0.0
+
+* Renamed gem from yaml2tmx to any2tmx.
+* Added support for JSON and Android XML.

data/Rakefile

@@ -0,0 +1,7 @@
+# encoding: UTF-8
+
+require 'rubygems' unless ENV['NO_RUBYGEMS']
+require 'rubygems/package_task'
+require 'bundler'
+
+Bundler::GemHelper.install_tasks

data/any2tmx.gemspec

@@ -0,0 +1,26 @@
+$:.unshift File.join(File.dirname(__FILE__), 'lib')
+require 'any2tmx/version'
+
+Gem::Specification.new do |s|
+  s.name     = "any2tmx"
+  s.version  = ::Any2Tmx::VERSION
+  s.authors  = ["Cameron Dutro"]
+  s.email    = ["camertron@gmail.com"]
+  s.homepage = "http://github.com/camertron/any2tmx"
+
+  s.description = s.summary = "A command-line tool to convert certain file types to the standard TMX format for translation memories."
+
+  s.platform = Gem::Platform::RUBY
+  s.has_rdoc = true
+
+  s.add_dependency 'xml-write-stream', '~> 1.0'
+  s.add_dependency 'nokogiri', '~> 1.6'
+  s.add_dependency 'htmlentities', '~> 4.3'
+
+  s.executables << 'yaml2tmx'
+  s.executables << 'json2tmx'
+  s.executables << 'android2tmx'
+
+  s.require_path = 'lib'
+  s.files = Dir["{lib,spec}/**/*", "Gemfile", "History.txt", "README.md", "Rakefile", "any2tmx.gemspec"]
+end

data/bin/android2tmx

@@ -0,0 +1,32 @@
+#! /usr/bin/env ruby
+
+require 'any2tmx'
+
+options = Any2Tmx::Options.new('android2tmx')
+
+unless options.valid?
+  puts options.errors.first
+  exit 1
+end
+
+if options.help?
+  options.print_help
+  exit 0
+end
+
+transform = Any2Tmx::Transforms::AndroidTransform.new(options)
+result = transform.result
+
+STDERR.write("Matched #{result.processed_count} of #{result.source_phrase_count} source phrases\n")
+
+stream = if options.output
+  File.open(options.output, 'w+')
+else
+  STDOUT
+end
+
+result.write(stream)
+
+if options.output
+  STDERR.write("Wrote #{options.output}\n")
+end

data/bin/json2tmx

@@ -0,0 +1,32 @@
+#! /usr/bin/env ruby
+
+require 'any2tmx'
+
+options = Any2Tmx::Options.new('json2tmx')
+
+unless options.valid?
+  puts options.errors.first
+  exit 1
+end
+
+if options.help?
+  options.print_help
+  exit 0
+end
+
+transform = Any2Tmx::Transforms::JsonTransform.new(options)
+result = transform.result
+
+STDERR.write("Matched #{result.processed_count} of #{result.source_phrase_count} source phrases\n")
+
+stream = if options.output
+  File.open(options.output, 'w+')
+else
+  STDOUT
+end
+
+result.write(stream)
+
+if options.output
+  STDERR.write("Wrote #{options.output}\n")
+end

data/bin/yaml2tmx

@@ -0,0 +1,32 @@
+#! /usr/bin/env ruby
+
+require 'any2tmx'
+
+options = Any2Tmx::Options.new('yaml2tmx')
+
+unless options.valid?
+  puts options.errors.first
+  exit 1
+end
+
+if options.help?
+  options.print_help
+  exit 0
+end
+
+transform = Any2Tmx::Transforms::YamlTransform.new(options)
+result = transform.result
+
+STDERR.write("Matched #{result.processed_count} of #{result.source_phrase_count} source phrases\n")
+
+stream = if options.output
+  File.open(options.output, 'w+')
+else
+  STDOUT
+end
+
+result.write(stream)
+
+if options.output
+  STDERR.write("Wrote #{options.output}\n")
+end

data/lib/any2tmx/android_xml_parser.rb

@@ -0,0 +1,146 @@
+require 'nokogiri'
+require 'htmlentities'
+
+# These classes have been adapted from:
+# https://github.com/rosette-proj/rosette-extractor-xml
+
+class HTMLEntities
+  MAPPINGS['android_xml'] = MAPPINGS['xhtml1'].dup.tap do |mappings|
+    mappings.delete('apos')
+  end
+
+  FLAVORS << 'android_xml'
+
+  class AndroidXmlDecoder < Decoder
+    def initialize
+      super('android_xml')
+    end
+  end
+end
+
+module Any2Tmx
+  class AndroidXmlParser
+    class << self
+      def parse(xml_content)
+        doc = parse_xml(xml_content)
+        result = {}.tap do |result|
+          collector = lambda { |text, key| result[key] = text }
+          each_string_entry(doc, &collector)
+          each_array_entry(doc, &collector)
+          each_plural_entry(doc, &collector)
+        end
+      end
+
+      protected
+
+      def parse_xml(xml_content)
+        Nokogiri::XML(xml_content) do |config|
+          config.options = Nokogiri::XML::ParseOptions::NONET
+        end
+      end
+
+      def each_string_entry(doc)
+        doc.xpath('//string').each do |node|
+          yield(
+            text_from(node),
+            name_from(node)
+          )
+        end
+      end
+
+      def each_array_entry(doc)
+        doc.xpath('//string-array').each do |array|
+          prefix = name_from(array)
+
+          array.xpath('item').each_with_index do |item, idx|
+            yield(
+              text_from(item),
+              "#{prefix}.#{idx}"
+            )
+          end
+        end
+      end
+
+      def each_plural_entry(doc)
+        doc.xpath('//plurals').each do |plurals|
+          prefix = name_from(plurals)
+
+          plurals.xpath('item').each do |item|
+            quantity = item.attributes['quantity'].value
+
+            yield(
+              text_from(item),
+              "#{prefix}.#{quantity}"
+            )
+          end
+        end
+      end
+
+      def text_from(node)
+        builder = Nokogiri::XML::Builder.new do |builder|
+          builder.root do
+            node.children.each do |child|
+              serialize(child, builder)
+            end
+          end
+        end
+
+        # safe to call `strip` after `to_xml` because any string that
+        # needs leading or trailing whitespace preserved should be wrapped
+        # in double quotes
+        unescape(
+          strip_enclosing_quotes(
+            builder.doc.xpath('/root/node()').to_xml.strip
+          )
+        )
+      end
+
+      def serialize(node, builder)
+        if node.text?
+          builder.text(unescape(node.text))
+        else
+          builder.send("#{node.name}_", node.attributes) do
+            node.children.each do |child|
+              serialize(child, builder)
+            end
+          end
+        end
+      end
+
+      def name_from(node)
+        if attribute = node.attributes['name']
+          attribute.value
+        end
+      end
+
+      def unescape(text)
+        text = text
+          .gsub("\\'", "'")
+          .gsub('\\"', '"')
+          .gsub("\\n", "\n")
+          .gsub("\\r", "\r")
+          .gsub("\\t", "\t")
+
+        coder.decode(text)
+      end
+
+      def coder
+        @coder ||= HTMLEntities::AndroidXmlDecoder.new
+      end
+
+      def strip_enclosing_quotes(text)
+        quote = case text[0]
+          when "'", '"'
+            text[0]
+        end
+
+        if quote
+          text.gsub(/\A#{quote}(.*)#{quote}\z/) { $1 }
+        else
+          text
+        end
+      end
+
+    end
+  end
+end

data/lib/any2tmx/options.rb

@@ -0,0 +1,120 @@
+require 'optparse'
+
+module Any2Tmx
+  class Options
+    attr_reader :errors
+
+    def initialize(executable_name)
+      @options = {}
+      @errors = []
+
+      OptionParser.new do |opts|
+        opts.banner = "Usage: #{executable_name} [options]"
+
+        opts.on('-s', '--source [file]:[locale]', 'File containing phrases in the source locale (locale appended with colon).') do |source|
+          file, locale = source.split(':')
+          @options[:source] = file
+          @options[:source_locale] = locale
+        end
+
+        opts.on('-t', '--target [file]:[locale]', 'File containing translations in the target locale (locale appended with colon).') do |target|
+          file, locale = target.split(':')
+          @options[:target] = file
+          @options[:target_locale] = locale
+        end
+
+        opts.on('-o', '--output [file]', 'The TMX output file to write. If not specified, output is printed to stdout.') do |output|
+          @options[:output] = output
+        end
+
+        opts.on('-h', '--help', 'Prints this help message.') do
+          @options[:help] = true
+        end
+
+        @opts = opts
+
+        # give derived classes the opportunity to add additional options
+        yield opts if block_given?
+      end.parse!
+    end
+
+    def valid?
+      errors.clear
+      validate
+      errors.empty?
+    end
+
+    def source
+      @options[:source]
+    end
+
+    def target
+      @options[:target]
+    end
+
+    def source_locale
+      @options[:source_locale]
+    end
+
+    def target_locale
+      @options[:target_locale]
+    end
+
+    def output
+      @options[:output]
+    end
+
+    def help?
+      @options[:help]
+    end
+
+    def print_help
+      puts @opts
+    end
+
+    # allow derived classes to add additional validation routines
+    def before_validate
+    end
+
+    def after_validate
+    end
+
+    private
+
+    def validate
+      unless help?
+        before_validate
+        validate_source
+        validate_target
+        validate_output
+        after_validate
+      end
+    end
+
+    def validate_source
+      unless File.exist?(source)
+        errors << 'Source file does not exist.'
+      end
+
+      unless source_locale
+        errors << 'Source locale not provided. Try running with the -h option for usage details.'
+      end
+    end
+
+    def validate_target
+      unless File.exist?(target)
+        errors << 'Target file does not exist.'
+      end
+
+      unless target_locale
+        errors << 'Target locale not provided. Try running with the -h option for usage details.'
+      end
+    end
+
+    def validate_output
+      unless File.exist?(File.dirname(target))
+        errors << 'Output path does not exist.'
+      end
+    end
+  end
+end

data/lib/any2tmx/phrase_set.rb

@@ -0,0 +1,14 @@
+module Any2Tmx
+  class PhraseSet
+    attr_reader :traversable, :locale
+
+    def initialize(traversable, locale)
+      @traversable = traversable
+      @locale = locale
+    end
+
+    def zip(other_set, &block)
+      traversable.zip(other_set.traversable, &block)
+    end
+  end
+end

data/lib/any2tmx/tmx_writer.rb

@@ -0,0 +1,34 @@
+require 'xml-write-stream'
+
+module Any2Tmx
+  class TmxWriter
+    class << self
+      def write(trans_map, source_locale, target_locale, io)
+        writer = XmlWriteStream.from_stream(io)
+        writer.open_tag('tmx', version: '1.4')
+        writer.open_single_line_tag('header', srclang: source_locale, datatype: 'plaintext', segtype: 'paragraph')
+        writer.close_tag
+
+        writer.open_tag('body')
+
+        trans_map.each_pair do |source_phrase, target_phrase|
+          writer.open_tag('tu')
+          writer.open_tag('tuv', 'xml:lang' => source_locale)
+          writer.open_single_line_tag('seg')
+          writer.write_text(source_phrase)
+          writer.close_tag  # seg
+          writer.close_tag  # tuv
+
+          writer.open_tag('tuv', 'xml:lang' => target_locale)
+          writer.open_single_line_tag('seg')
+          writer.write_text(target_phrase)
+          writer.close_tag  # seg
+          writer.close_tag  # tuv
+          writer.close_tag  # tu
+        end
+
+        writer.close
+      end
+    end
+  end
+end

data/lib/any2tmx/transforms/android_transform.rb

@@ -0,0 +1,13 @@
+module Any2Tmx
+  module Transforms
+    class AndroidTransform < Transform
+      private
+
+      def load(file, locale)
+        phrases = Any2Tmx::AndroidXmlParser.parse(File.read(file))
+        traversable = Any2Tmx::Traversable.new(phrases)
+        Any2Tmx::PhraseSet.new(traversable, locale)
+      end
+    end
+  end
+end

data/lib/any2tmx/transforms/json_transform.rb

@@ -0,0 +1,15 @@
+require 'json'
+
+module Any2Tmx
+  module Transforms
+    class JsonTransform < Transform
+      private
+
+      def load(file, locale)
+        phrases = JSON.parse(File.read(file))
+        traversable = Any2Tmx::Traversable.new(phrases)
+        Any2Tmx::PhraseSet.new(traversable, locale)
+      end
+    end
+  end
+end

data/lib/any2tmx/transforms/result.rb

@@ -0,0 +1,24 @@
+module Any2Tmx
+  module Transforms
+    class Result
+      attr_reader :source, :target, :collection, :processed_count
+
+      def initialize(source, target, collection, processed_count)
+        @source = source
+        @target = target
+        @collection = collection
+        @processed_count = processed_count
+      end
+
+      def source_phrase_count
+        source.traversable.size
+      end
+
+      def write(io)
+        Any2Tmx::TmxWriter.write(
+          collection, source.locale, target.locale, io
+        )
+      end
+    end
+  end
+end

data/lib/any2tmx/transforms/transform.rb

@@ -0,0 +1,26 @@
+module Any2Tmx
+  module Transforms
+    class Transform
+      attr_reader :options
+
+      def initialize(options)
+        @options = options
+      end
+
+      def result
+        source = load(options.source, options.source_locale)
+        target = load(options.target, options.target_locale)
+        count = 0
+        zipped = source.zip(target) { count += 1 }
+        Result.new(source, target, zipped, count)
+      end
+
+      private
+
+      def load(file, locale)
+        raise NotImplementedError,
+          "#{__method__} must be defined in derived classes"
+      end
+    end
+  end
+end

data/lib/any2tmx/transforms/yaml_transform.rb

@@ -0,0 +1,20 @@
+require 'yaml'
+
+module Any2Tmx
+  module Transforms
+    class YamlTransform < Transform
+      private
+
+      def load(file, locale)
+        phrases = YAML.load_file(file)
+
+        if phrases.include?(locale)
+          phrases = phrases[locale]
+        end
+
+        traversable = Any2Tmx::Traversable.new(phrases)
+        Any2Tmx::PhraseSet.new(traversable, locale)
+      end
+    end
+  end
+end

data/lib/any2tmx/transforms.rb

@@ -0,0 +1,9 @@
+module Any2Tmx
+  module Transforms
+    autoload :AndroidTransform, 'any2tmx/transforms/android_transform'
+    autoload :JsonTransform,    'any2tmx/transforms/json_transform'
+    autoload :Transform,        'any2tmx/transforms/transform'
+    autoload :Result,           'any2tmx/transforms/result'
+    autoload :YamlTransform,    'any2tmx/transforms/yaml_transform'
+  end
+end

data/lib/any2tmx/traversable.rb

@@ -0,0 +1,62 @@
+module Any2Tmx
+  class Traversable
+    attr_reader :collection
+
+    def initialize(collection)
+      @collection = collection
+    end
+
+    def each_entry(&block)
+      if block_given?
+        each_entry_helper(collection, [], &block)
+      else
+        to_enum(__method__)
+      end
+    end
+
+    def size
+      each_entry.inject(0) { |ret, _| ret + 1 }
+    end
+
+    def dig(path)
+      path.inject(:start) do |ret, seg|
+        if ret == :start
+          collection[seg]
+        elsif ret
+          if seg.is_a?(Numeric) && ret.is_a?(Array)
+            ret[seg]  # array index case
+          elsif seg.is_a?(String) && ret.is_a?(Hash)
+            ret[seg]  # hash key case
+          end
+        end
+      end
+    end
+
+    def zip(other_traversable)
+      each_entry.each_with_object({}) do |(entry, path), ret|
+        other_entry = other_traversable.dig(path)
+        ret[entry] = other_entry if other_entry
+        yield if block_given? && other_entry
+      end
+    end
+
+    private
+
+    def each_entry_helper(coll, path, &block)
+      case coll
+        when Hash
+          coll.each_pair do |key, value|
+            each_entry_helper(value, path + [key], &block)
+          end
+
+        when Array
+          coll.each_with_index do |element, idx|
+            each_entry_helper(element, path + [idx], &block)
+          end
+
+        when String
+          yield coll, path
+      end
+    end
+  end
+end

data/lib/any2tmx/version.rb

@@ -0,0 +1,3 @@
+module Any2Tmx
+  VERSION = '1.0.0'
+end

data/lib/any2tmx.rb

@@ -0,0 +1,8 @@
+module Any2Tmx
+  autoload :AndroidXmlParser, 'any2tmx/android_xml_parser'
+  autoload :Options,          'any2tmx/options'
+  autoload :PhraseSet,        'any2tmx/phrase_set'
+  autoload :TmxWriter,        'any2tmx/tmx_writer'
+  autoload :Transforms,       'any2tmx/transforms'
+  autoload :Traversable,      'any2tmx/traversable'
+end

metadata

@@ -0,0 +1,111 @@
+--- !ruby/object:Gem::Specification
+name: any2tmx
+version: !ruby/object:Gem::Version
+  version: 1.0.0
+platform: ruby
+authors:
+- Cameron Dutro
+autorequire: 
+bindir: bin
+cert_chain: []
+date: 2016-01-28 00:00:00.000000000 Z
+dependencies:
+- !ruby/object:Gem::Dependency
+  name: xml-write-stream
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '1.0'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '1.0'
+- !ruby/object:Gem::Dependency
+  name: nokogiri
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '1.6'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '1.6'
+- !ruby/object:Gem::Dependency
+  name: htmlentities
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '4.3'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '4.3'
+description: A command-line tool to convert certain file types to the standard TMX
+  format for translation memories.
+email:
+- camertron@gmail.com
+executables:
+- yaml2tmx
+- json2tmx
+- android2tmx
+extensions: []
+extra_rdoc_files: []
+files:
+- Gemfile
+- History.txt
+- Rakefile
+- any2tmx.gemspec
+- bin/android2tmx
+- bin/json2tmx
+- bin/yaml2tmx
+- lib/any2tmx.rb
+- lib/any2tmx/android_xml_parser.rb
+- lib/any2tmx/options.rb
+- lib/any2tmx/phrase_set.rb
+- lib/any2tmx/tmx_writer.rb
+- lib/any2tmx/transforms.rb
+- lib/any2tmx/transforms/android_transform.rb
+- lib/any2tmx/transforms/json_transform.rb
+- lib/any2tmx/transforms/result.rb
+- lib/any2tmx/transforms/transform.rb
+- lib/any2tmx/transforms/yaml_transform.rb
+- lib/any2tmx/traversable.rb
+- lib/any2tmx/version.rb
+homepage: http://github.com/camertron/any2tmx
+licenses: []
+metadata: {}
+post_install_message: 
+rdoc_options: []
+require_paths:
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - ">="
+    - !ruby/object:Gem::Version
+      version: '0'
+required_rubygems_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - ">="
+    - !ruby/object:Gem::Version
+      version: '0'
+requirements: []
+rubyforge_project: 
+rubygems_version: 2.2.3
+signing_key: 
+specification_version: 4
+summary: A command-line tool to convert certain file types to the standard TMX format
+  for translation memories.
+test_files: []
+has_rdoc: true