get_pomo 0.6.0

data/.gitignore

@@ -0,0 +1 @@
+pkg

data/README.markdown

@@ -0,0 +1,47 @@
+A simple and extendable .mo and .po file parser/generator.  
+
+Advanteges over original [mo](http://github.com/mutoh/gettext/blob/abf96713327cc4c5d35f0a772f3b75ff4819450c/lib/gettext/mofile.rb) / [po](http://github.com/mutoh/gettext/blob/abf96713327cc4c5d35f0a772f3b75ff4819450c/lib/gettext/poparser.rb)-parser:
+
+ - simple architecture + easy to extend/modify
+ - emtpy msgstr translations are read
+ - comments are included
+ - fuzzy can be set/unset
+ - multiple translations can be combined in a new po file(with comments and fuzzy and ...)
+ - po files can be written from any kind of input
+ - easy mo-file handling/merging
+ - po/mo file handling is identical, if you know one, you know both
+
+Setup
+=====
+    sudo gem install get_pomo -s http://gemcutter.org
+
+###Static interface
+    #parse po files
+    translations = GetPomo::PoFile.parse(File.read('xxx.po')) + GetPomo::PoFile.parse(File.read('yyy.po'))
+
+    #and use the data...
+    msgids = translations.reject{|t|t.plural? or t.fuzzy?}.map(&:msgid)
+
+    #or write a new po file (unique by msgid)...
+    File.open('xxx.po','w){|f|f.print(GetPomo::PoFile.to_text(translations))}
+
+
+###Instance interface
+    p = GetPomo::PoFile.new
+    p.add_translations_from_text(File.read('...'))
+    ...
+    p.translations
+    p.to_text
+
+`GetPomo::MoFile` behaves identical.
+
+TODO
+====
+ - extracting of version/pluralisation_rule/plurals/translator... (from msgid "")
+ - the vendor/mofile is really complex, maybe it can be refactored (also some parts are not needed)
+
+Author
+======
+[Michael Grosser](http://pragmatig.wordpress.com)  
+grosser.michael@gmail.com  
+Hereby placed under public domain, do what you want, just do not hold me accountable...  

data/Rakefile

@@ -0,0 +1,19 @@
+task :default => :spec
+require 'spec/rake/spectask'
+Spec::Rake::SpecTask.new {|t| t.spec_opts = ['--color']}
+
+begin
+  project_name = 'get_pomo'
+  require 'jeweler'
+  Jeweler::Tasks.new do |gem|
+    gem.name = project_name
+    gem.summary = "Ruby/Gettext: A .po and .mo file parser/generator"
+    gem.email = "grosser.michael@gmail.com"
+    gem.homepage = "http://github.com/grosser/#{project_name}"
+    gem.authors = ["Michael Grosser"]
+  end
+
+  Jeweler::GemcutterTasks.new
+rescue LoadError
+  puts "Jeweler, or one of its dependencies, is not available. Install it with: sudo gem install technicalpickles-jeweler -s http://gems.github.com"
+end

data/VERSION

@@ -0,0 +1 @@
+0.6.0

data/get_pomo.gemspec

@@ -0,0 +1,66 @@
+# Generated by jeweler
+# DO NOT EDIT THIS FILE DIRECTLY
+# Instead, edit Jeweler::Tasks in Rakefile, and run the gemspec command
+# -*- encoding: utf-8 -*-
+
+Gem::Specification.new do |s|
+  s.name = %q{get_pomo}
+  s.version = "0.6.0"
+
+  s.required_rubygems_version = Gem::Requirement.new(">= 0") if s.respond_to? :required_rubygems_version=
+  s.authors = ["Michael Grosser"]
+  s.date = %q{2009-11-20}
+  s.email = %q{grosser.michael@gmail.com}
+  s.extra_rdoc_files = [
+    "README.markdown"
+  ]
+  s.files = [
+    ".gitignore",
+     "README.markdown",
+     "Rakefile",
+     "VERSION",
+     "get_pomo.gemspec",
+     "lib/get_pomo.rb",
+     "lib/get_pomo/mo_file.rb",
+     "lib/get_pomo/po_file.rb",
+     "lib/get_pomo/translation.rb",
+     "prototype_treetop/po.treetop",
+     "prototype_treetop/test.rb",
+     "spec/files/complex.mo",
+     "spec/files/empty.mo",
+     "spec/files/plural.mo",
+     "spec/files/singular.mo",
+     "spec/files/singular_2.mo",
+     "spec/pomo/mo_file_spec.rb",
+     "spec/pomo/po_file_spec.rb",
+     "spec/pomo/translation_spec.rb",
+     "spec/pomo_spec.rb",
+     "spec/spec_helper.rb",
+     "vendor/README.rdoc",
+     "vendor/iconv.rb",
+     "vendor/mofile.rb"
+  ]
+  s.homepage = %q{http://github.com/grosser/get_pomo}
+  s.rdoc_options = ["--charset=UTF-8"]
+  s.require_paths = ["lib"]
+  s.rubygems_version = %q{1.3.5}
+  s.summary = %q{Ruby/Gettext: A .po and .mo file parser/generator}
+  s.test_files = [
+    "spec/spec_helper.rb",
+     "spec/pomo_spec.rb",
+     "spec/pomo/translation_spec.rb",
+     "spec/pomo/mo_file_spec.rb",
+     "spec/pomo/po_file_spec.rb"
+  ]
+
+  if s.respond_to? :specification_version then
+    current_version = Gem::Specification::CURRENT_SPECIFICATION_VERSION
+    s.specification_version = 3
+
+    if Gem::Version.new(Gem::RubyGemsVersion) >= Gem::Version.new('1.2.0') then
+    else
+    end
+  else
+  end
+end
+

data/lib/get_pomo/mo_file.rb

@@ -0,0 +1,62 @@
+require 'get_pomo/translation'
+require File.join(File.dirname(__FILE__),'..','..','vendor','mofile')
+
+module GetPomo
+  class MoFile
+    PLURAL_SEPERATOR = "\000"
+
+    def self.parse(text)
+      MoFile.new.add_translations_from_text(text)
+    end
+
+    def self.to_text(translations)
+      m = MoFile.new(:translations=>translations)
+      m.to_text
+    end
+
+    attr_reader :translations
+
+    def initialize(options = {})
+      @translations = options[:translations] || []
+    end
+
+    def add_translations_from_text(text)
+      text = StringIO.new(text)
+      @translations += GetPomo::GetText::MOFile.open(text, "UTF-8").map do |msgid,msgstr|
+        translation = Translation.new
+        if plural? msgid or plural? msgstr
+          translation.msgid = split_plural(msgid)
+          translation.msgstr = split_plural(msgstr)
+        else
+          translation.msgid = msgid
+          translation.msgstr = msgstr
+        end
+        translation
+      end
+    end
+
+    def to_text
+      m = GetPomo::GetText::MOFile.new
+      GetPomo.unique_translations(translations).each {|t| m[plural_to_string(t.msgid)] = plural_to_string(t.msgstr)}
+
+      io = StringIO.new
+      m.save_to_stream io
+      io.rewind
+      io.read
+    end
+
+    private
+
+    def plural_to_string(plural_or_singular)
+      [*plural_or_singular] * PLURAL_SEPERATOR
+    end
+
+    def plural? string
+      string.include? PLURAL_SEPERATOR
+    end
+
+    def split_plural string
+      string.split PLURAL_SEPERATOR
+    end
+  end
+end

data/lib/get_pomo/po_file.rb

@@ -0,0 +1,111 @@
+require 'get_pomo/translation'
+
+module GetPomo
+  class PoFile
+    def self.parse(text)
+      PoFile.new.add_translations_from_text(text)
+    end
+
+    def self.to_text(translations)
+      p = PoFile.new(:translations=>translations)
+      p.to_text
+    end
+
+    attr_reader :translations
+
+    def initialize(options = {})
+      @translations = options[:translations] || []
+    end
+
+    #the text is split into lines and then converted into logical translations
+    #each translation consists of comments(that come before a translation)
+    #and a msgid / msgstr
+    def add_translations_from_text(text)
+      start_new_translation
+      text.split(/$/).each_with_index do |line,index|
+        @line_number = index + 1
+        next if line.empty?
+        if method_call? line
+          parse_method_call line
+        elsif comment? line
+          add_comment line
+        else
+          add_string line
+        end
+      end
+      start_new_translation #instance_variable has to be overwritten or errors can occur on next add
+      translations
+    end
+
+    def to_text
+      GetPomo.unique_translations(translations).map {|translation|
+        comment = translation.comment.to_s.split(/\n|\r\n/).map{|line|"##{line}\n"}*''
+        msgid_and_msgstr = if translation.plural?
+          msgids =
+          %Q(msgid "#{translation.msgid[0]}"\n)+
+          %Q(msgid_plural "#{translation.msgid[1]}"\n)
+
+          msgstrs = []
+          translation.msgstr.each_with_index do |msgstr,index|
+            msgstrs << %Q(msgstr[#{index}] "#{msgstr}")
+          end
+
+          msgids + (msgstrs*"\n")
+        else
+          %Q(msgid "#{translation.msgid}"\n)+
+          %Q(msgstr "#{translation.msgstr}")
+        end
+
+        comment + msgid_and_msgstr
+      } * "\n\n"
+    end
+
+    private
+
+    #e.g. # fuzzy
+    def comment?(line)
+      line =~ /^\s*#/
+    end
+
+    def add_comment(line)
+      start_new_translation if translation_complete?
+      @current_translation.add_text(line.strip.sub('#','')+"\n",:to=>:comment)
+    end
+
+    #msgid "hello"
+    def method_call?(line)
+      line =~ /^\s*[a-z]/
+    end
+
+    #msgid "hello" -> method call msgid + add string "hello"
+    def parse_method_call(line)
+      method, string = line.match(/^\s*([a-z0-9_\[\]]+)(.*)/)[1..2]
+      raise "no method found" unless method
+
+      start_new_translation if method == 'msgid' and translation_complete?
+      @last_method = method.to_sym
+      add_string(string)
+    end
+
+    #"hello" -> hello
+    def add_string(string)
+      return if string.strip.empty?
+      raise "not string format: #{string.inspect} on line #{@line_number}" unless string.strip =~ /^['"](.*)['"]$/
+      @current_translation.add_text($1,:to=>@last_method)
+    end
+
+    def translation_complete?
+      return false unless @current_translation
+      @current_translation.complete?
+    end
+  
+    def store_translation
+      @translations += [@current_translation] if @current_translation.complete?
+    end
+
+    def start_new_translation
+      store_translation if translation_complete?
+      @current_translation = Translation.new
+    end
+  end
+end

data/lib/get_pomo/translation.rb

@@ -0,0 +1,44 @@
+module GetPomo
+  class Translation
+    FUZZY_REGEX = /^\s*fuzzy\s*$/
+    attr_accessor :msgid, :msgstr, :comment
+
+    def add_text(text,options)
+      to = options[:to]
+      if to.to_sym == :msgid_plural
+        self.msgid = [msgid] unless msgid.is_a? Array
+        msgid[1] = msgid[1].to_s + text
+      elsif to.to_s =~ /^msgstr\[(\d)\]$/
+        self.msgstr ||= []
+        msgstr[$1.to_i] = msgstr[$1.to_i].to_s + text
+      else
+        #simple form
+        send("#{to}=",send(to).to_s+text)
+      end
+    end
+
+    def to_hash
+      {:msgid=>msgid,:msgstr=>msgstr,:comment=>comment}.reject{|k,value|value.nil?}
+    end
+
+    def complete?
+      not msgid.nil? and not msgstr.nil?
+    end
+
+    def fuzzy?
+      comment =~ FUZZY_REGEX
+    end
+
+    def fuzzy=(value)
+      if value and not fuzzy?
+        add_text "\nfuzzy", :to=>:comment
+      else
+        self.comment = comment.to_s.split(/$/).reject{|line|line=~FUZZY_REGEX}.join("\n")
+      end
+    end
+
+    def plural?
+      msgid.is_a? Array or msgstr.is_a? Array
+    end
+  end
+end

data/lib/get_pomo.rb

@@ -0,0 +1,12 @@
+require 'get_pomo/po_file'
+module GetPomo
+  VERSION = File.read( File.join(File.dirname(__FILE__),'..','VERSION') ).strip
+  
+  extend self
+  
+  def self.unique_translations(translations)
+    last_seen_at_index = {}
+    translations.each_with_index {|translation,index|last_seen_at_index[translation.msgid]=index}
+    last_seen_at_index.values.sort.map{|index| translations[index]}
+  end
+end

data/prototype_treetop/po.treetop

@@ -0,0 +1,25 @@
+grammar Po
+  rule translation
+    msgid whitespace string
+  end
+
+  rule string
+    string whitespace string
+  end
+
+  rule quote
+    '"'
+  end
+
+  rule text
+    [^"]*
+  end
+
+  rule whitespace
+    [ \t\n\r]+
+  end
+
+  rule msgid
+    "msgid"
+  end
+end

data/prototype_treetop/test.rb

@@ -0,0 +1,6 @@
+require 'rubygems'
+require 'treetop'
+
+Treetop.load "po"
+parser = PoParser.new
+puts parser.parse(%Q(msgid "xxx" "yyy"))

data/spec/pomo/mo_file_spec.rb

@@ -0,0 +1,58 @@
+require File.expand_path("../spec_helper", File.dirname(__FILE__))
+require 'get_pomo/mo_file'
+
+include GetPomo
+describe GetPomo::MoFile do
+  it "parses empty mo file" do
+    MoFile.parse(File.read('spec/files/empty.mo')).should == []
+  end
+
+  it "parses empty strings" do
+    MoFile.parse(File.read('spec/files/empty.mo')).should == []
+  end
+
+  it "reads singulars" do
+    t = MoFile.parse(File.read('spec/files/singular.mo'))[0]
+    t.to_hash.should == {:msgid=>'Back',:msgstr=>'Zurück'}
+  end
+
+  it "reads plurals" do
+    t = MoFile.parse(File.read('spec/files/plural.mo'))[0]
+    t.to_hash.should == {:msgid=>['Axis','Axis'],:msgstr=>['Achse','Achsen']}
+  end
+
+  describe 'instance methods' do
+    it "combines multiple translations" do
+      m = MoFile.new
+      m.add_translations_from_text(File.read('spec/files/plural.mo'))
+      m.add_translations_from_text(File.read('spec/files/singular.mo'))
+      m.should have(2).translations
+      m.translations[0].msgid.should_not == m.translations[1].msgid
+    end
+
+    it "can be initialized with translations" do
+      m = MoFile.new(:translations=>['x'])
+      m.translations.should == ['x']
+    end
+
+    it "does not generate duplicate translations" do
+      second_version = File.read('spec/files/singular_2.mo')
+      m = MoFile.new
+      m.add_translations_from_text(File.read('spec/files/singular.mo'))
+      m.add_translations_from_text(second_version)
+      m.to_text.should == second_version
+    end
+  end
+
+  it "reads metadata" do
+    meta = MoFile.parse(File.read('spec/files/complex.mo')).detect {|t|t.msgid == ''}
+    meta.msgstr.should_not be_empty
+  end
+
+  describe :to_text do
+    it "writes singulars" do
+      text = File.read('spec/files/singular.mo')
+      MoFile.to_text(MoFile.parse(text)).should == text
+    end
+  end
+end

data/spec/pomo/po_file_spec.rb

@@ -0,0 +1,117 @@
+require File.expand_path("../spec_helper", File.dirname(__FILE__))
+
+include GetPomo
+describe GetPomo::PoFile do
+  describe :parse do
+    it "parses nothing" do
+      PoFile.parse("").should be_empty
+    end
+
+    it "parses a simple msgid and msgstr" do
+      t = PoFile.parse(%Q(msgid "xxx"\nmsgstr "yyy"))
+      t[0].to_hash.should == {:msgid=>'xxx',:msgstr=>'yyy'}
+    end
+
+    it "parses a simple msgid and msg with additional whitespace" do
+      t = PoFile.parse(%Q(  msgid    "xxx"   \n   msgstr    "yyy"   ))
+      t[0].to_hash.should == {:msgid=>'xxx',:msgstr=>'yyy'}
+    end
+
+    it "parses an empty msgid with text (gettext meta data)" do
+      t = PoFile.parse(%Q(msgid ""\nmsgstr "PLURAL FORMS"))
+      t[0].to_hash.should == {:msgid=>'',:msgstr=>'PLURAL FORMS'}
+    end
+
+    it "parses a multiline msgid/msgstr" do
+      t = PoFile.parse(%Q(msgid "xxx"\n"aaa"\n\n\nmsgstr ""\n"bbb"))
+      t[0].to_hash.should == {:msgid=>'xxxaaa',:msgstr=>'bbb'}
+    end
+
+    it "parses simple comments" do
+      t = PoFile.parse(%Q(#test\nmsgid "xxx"\nmsgstr "yyy"))
+      t[0].to_hash.should == {:msgid=>'xxx',:msgstr=>'yyy',:comment=>"test\n"}
+    end
+
+    it "parses comments above msgstr" do
+      t = PoFile.parse(%Q(#test\nmsgid "xxx"\n#another\nmsgstr "yyy"))
+      t[0].to_hash.should == {:msgid=>'xxx',:msgstr=>'yyy',:comment=>"test\nanother\n"}
+    end
+  end
+
+  describe "instance interface" do
+    it "adds two different translations" do
+      p = PoFile.new
+      p.add_translations_from_text(%Q(msgid "xxx"\nmsgstr "yyy"))
+      p.add_translations_from_text(%Q(msgid "aaa"\nmsgstr "yyy"))
+      p.translations[1].to_hash.should == {:msgid=>'aaa',:msgstr=>'yyy'}
+    end
+
+    it "can be initialized with translations" do
+      p = PoFile.new(:translations=>['xxx'])
+      p.translations[0].should == 'xxx'
+    end
+
+    it "can be converted to text" do
+      text = %Q(msgid "xxx"\nmsgstr "aaa")
+      p = PoFile.new
+      p.add_translations_from_text(text)
+      p.to_text.should == text
+    end
+
+    it "keeps uniqueness when converting to_text" do
+      text = %Q(msgid "xxx"\nmsgstr "aaa")
+      p = PoFile.new
+      p.add_translations_from_text(%Q(msgid "xxx"\nmsgstr "yyy"))
+      p.add_translations_from_text(text)
+      p.to_text.should == text
+    end
+  end
+
+  it "adds plural translations" do
+    t = PoFile.parse(%Q(msgid "singular"\nmsgid_plural "plural"\nmsgstr[0] "one"\nmsgstr[1] "many"))
+    t[0].to_hash.should == {:msgid=>['singular','plural'],:msgstr=>['one','many']}
+  end
+
+  it "does not fail on empty string" do
+    PoFile.parse(%Q(\n\n\n\n\n))
+  end
+
+  it "shows line number for invalid strings" do
+    begin
+      PoFile.parse(%Q(\n\n\n\n\nmsgstr "))
+      flunk
+    rescue Exception => e
+      e.to_s.should =~ /line 5/
+    end
+  end
+
+  describe :to_text do
+    it "is empty when not translations where added" do
+      PoFile.to_text([]).should == ""
+    end
+    
+    it "preserves simple syntax" do
+      text = %Q(msgid "x"\nmsgstr "y")
+      PoFile.to_text(PoFile.parse(text)).should == text
+    end
+
+    it "adds comments" do
+      t = GetPomo::Translation.new
+      t.msgid = 'a'
+      t.msgstr = 'b'
+      t.add_text("c\n",:to=>:comment)
+      t.add_text("d\n",:to=>:comment)
+      PoFile.to_text([t]).should == %Q(#c\n#d\nmsgid "a"\nmsgstr "b")
+    end
+
+    it "uses plural notation" do
+      text = %Q(#awesome\nmsgid "one"\nmsgid_plural "many"\nmsgstr[0] "1"\nmsgstr[1] "n")
+      PoFile.to_text(PoFile.parse(text)).should == text
+    end
+
+    it "only uses the latest of identicals msgids" do
+      text = %Q(msgid "one"\nmsgstr "1"\nmsgid "one"\nmsgstr "001")
+      PoFile.to_text(PoFile.parse(text)).should ==  %Q(msgid "one"\nmsgstr "001")
+    end
+  end
+end