yury-typograph 0.1.1

data/.gitignore

@@ -0,0 +1,2 @@
+.idea
+pkg

data/README

@@ -0,0 +1 @@
+Ruby port of http://emuravjev.ru/works/tg php typograph

data/Rakefile

@@ -0,0 +1,13 @@
+begin
+  require 'jeweler'
+  Jeweler::Tasks.new do |gemspec|
+    gemspec.name = "typograph"
+    gemspec.summary = "ruby html typograph"
+    gemspec.description = "apply some basic typography rules to text"
+    gemspec.email = "yury.korolev@gmail.com"
+    gemspec.homepage = "http://github.com/yury/typograph"
+    gemspec.authors = ["Yury Korolev"]
+  end
+rescue LoadError
+  puts "Jeweler not available. Install it with: sudo gem install technicalpickles-jeweler -s http://gems.github.com"
+end

data/VERSION

@@ -0,0 +1 @@
+0.1.1

data/lib/typograph.rb

@@ -0,0 +1,23 @@
+module Typograph
+  autoload :Rules, 'typograph/rules'
+
+  @@rules = [Rules::QuoteRules.new,
+             Rules::DashRules.new,
+             Rules::PunchmarkRules.new,
+             Rules::MathRules.new,
+             Rules::EtcRules.new] 
+
+  def self.run! text
+    for rule in @@rules
+      next if rule.disabled
+      text = rule.parse text
+    end
+    text
+  end
+
+  def self.run text
+    self.run! text.dup
+  end
+end
+
+require 'typograph/string_sub_each'

data/lib/typograph/rules.rb

@@ -0,0 +1,64 @@
+module Typograph
+  module Rules
+
+    autoload :DashRules, 'typograph/rules/dash_rules'
+    autoload :EtcRules, 'typograph/rules/etc_rules'
+    autoload :MathRules, 'typograph/rules/math_rules'
+    autoload :PunchmarkRules, 'typograph/rules/punchmark_rules'
+    autoload :QuoteRules, 'typograph/rules/quote_rules'
+
+    # base class for all rules in typograph
+    class Rule
+
+      NBSP = " "
+      QUOTE_FIRS_OPEN = '«';
+      QUOTE_FIRS_CLOSE = '»';
+      QUOTE_CRAWSE_OPEN = '„';
+      QUOTE_CRAWSE_CLOSE = '“';
+
+      attr_accessor :disabled
+
+      def initialize
+        @disabled = false
+        @rules = []
+      end
+
+      def add name, opts = {}
+        @rules << {:name => name, :opts => opts}
+      end
+
+      def pre_parse text
+        text
+      end
+
+      def post_parse text
+        text
+      end
+
+      def parse text
+        return text if self.disabled
+
+        text = self.pre_parse text
+
+        @rules.each do |rule|
+          opts = rule[:opts]
+          next if opts[:disabled]
+          replacement = opts[:replacement]
+          if replacement.class == Proc
+            text.sub_each! opts[:pattern], &replacement
+          else
+            text.gsub! opts[:pattern], opts[:replacement]
+          end
+        end
+
+        self.post_parse text
+      end
+
+      def self.run text
+        t = self.new
+        t.parse text
+      end
+
+    end
+  end
+end

data/lib/typograph/rules/dash_rules.rb

@@ -0,0 +1,41 @@
+module Typograph
+  module Rules
+
+    class DashRules < Rule
+
+      def initialize
+        super
+
+        add :m_dash,
+            :pattern => /([a-zа-я0-9]+|,|:|\)|&raquo;|\|")(\040|\t)(\-|&mdash;)(\s|$|<)/u,
+            :replacement => '\1&nbsp;&mdash;\4'
+
+        add :m_dash_2,
+            :pattern => /(\n|\r|^|>)(\-|&mdash;)(\t|\040)/,
+            :replacement => '\1&mdash;&nbsp;'
+
+        add :m_dash_3,
+            :pattern => /(\.|!|\?|&hellip;)(\040|\t|&nbsp;)(\-|&mdash;)(\040|\t|&nbsp;)/,
+            :replacement => '\1 &mdash;&nbsp;'
+
+        add :years,
+            :pattern => /(с|по|период|середины|начала|начало|конца|конец|половины|в|между)(\s+|&nbsp;)([\d]{4})(-)([\d]{4})(г|гг)?/eui,
+            :replacement => lambda {|m| "#{m[1]}#{m[2]}#{build_years(m[3],m[5],m[4])}#{m[6]}"  }
+
+        add :iz_za_pod,
+            :pattern => /(\s|&nbsp;|>)(из)(\040|\t|&nbsp;)\-?(за|под)([\.,!\?:;]|\040|&nbsp;)/uie,
+            :replacement => lambda {|m| "#{m[1] == NBSP ? " " : m[1]}#{m[2]}-#{m[4]}#{m[5] == NBSP ? " " : m[5]}" }
+
+        add :to_libo_nibud,
+            :pattern => /(\s|^|&nbsp;|>)(кто|кем|когда|зачем|почему|как|что|чем|где|чего|кого)\-?(\040|\t|&nbsp;)\-?(то|либо|нибудь)([\.,!?;]|\040|&nbsp;|$)/ui,
+            :replacement => lambda {|m| "#{m[1] == NBSP ? " " : m[1]}#{m[2]}-#{m[4]}#{m[5] == NBSP ? " " : m[5]}" }
+      end
+
+      def build_years start_date, end_date, sep
+        start_date = start_date.to_i
+        end_date = end_date.to_i
+        start_date >= end_date ? "#{start_date}#{sep}#{end_date}" : "#{start_date}&ndash;#{end_date}"
+      end
+    end
+  end
+end

data/lib/typograph/rules/etc_rules.rb

@@ -0,0 +1,31 @@
+module Typograph
+  module Rules
+
+    class EtcRules < Rule
+
+      def initialize
+        super
+        add :tm_replace,
+            :pattern => /([\040\t])?\(tm\)/i,
+            :replacement => '&trade;'
+
+        # TODO: encrypted tags
+        #add :r_sign_replace,
+        #    :pattern => /\(r\)/ei
+
+
+        add :copy_replace,
+            :pattern => /\((c|с)\)\s+/iu,
+            :replacement => '&copy;&nbsp;'
+
+        add :acute_accent,
+            :pattern => /(у|е|ы|а|о|э|я|и|ю|ё)`(\w)/iu,
+            :replacement => '\1&#769;\2'
+
+        #add :auto_links
+
+      end
+
+    end
+  end
+end

data/lib/typograph/rules/math_rules.rb

@@ -0,0 +1,33 @@
+module Typograph
+  module Rules
+
+    class MathRules < Rule
+
+      def initialize
+        super
+
+        add :auto_times_x,
+            :pattern => /(\d+)(\040*)(x|х)(\040*)(\d+)/u,
+            :replacement => '\1&times;\5'
+
+        add :frac_1, # replaces 1/2 and 1/4
+            :pattern => /(\D)1\/(2|4)(\D)/,
+            :replacement => '\1&frac1\2;\3'
+
+        add :frac_3, # replaces 3/4
+            :pattern => /(\D)3\/4(\D)/,
+            :replacement => '\1&frac34;\2'
+
+        add :not_equal,
+            :pattern => /!=/,
+            :replacement => '&#8800;'
+
+        add :plus_minus,
+            :pattern => /\+-/,
+            :replacement => '&plusmn;'
+
+      end
+
+    end
+  end
+end

data/lib/typograph/rules/punchmark_rules.rb

@@ -0,0 +1,45 @@
+module Typograph
+  module Rules
+
+    class PunchmarkRules < Rule
+
+      def initialize
+        super
+
+        add :auto_comma,
+            :pattern => /([a-zа-я])(\s|&nbsp;)(но|а)(\s|&nbsp;)/iu,
+            :replacement =>  '\1,\2\3\4'
+
+        add :punctuation_marks_limit,
+            :pattern => /([!.?]){4,}/,
+            :replacement => '\1\1\1'
+
+        add :punctuation_marks_base_limit,
+            :pattern => /([,:;]){2,}/,
+            :replacement => '\1'
+
+        add :hellip,
+            :pattern => /\.\.+/,
+            :replacement => '&hellip;'
+
+        add :eng_apostrophe,
+            :pattern => /(\s|^|>)([a-z]{2,})'([a-z]+)/i,
+            :replacement => '\1\2&rsquo;\3'
+
+        add :fix_pmarks,
+            :pattern => /([a-zа-я0-9])(!|\.|\?){2}(\s|$|<)/iu,
+            :replacement => '\1\2\3'
+
+        add :fix_l_bracket,
+            :pattern => /(\()(\040|\t)+/,
+            :replacement => '\1'
+
+        add :fix_r_bracket,
+            :pattern => /(\040|\t)+(\))/,
+            :replacement => '\2'
+
+      end
+
+    end
+  end
+end

data/lib/typograph/rules/quote_rules.rb

@@ -0,0 +1,24 @@
+module Typograph
+  module Rules
+
+    class QuoteRules < Rule
+
+      def initialize
+        super
+
+        add :quotes_outside_a,
+            :pattern => /(<%%__.+?>)"(.+?)"(<\/%%__.+?>)/s,
+            :replacement => '"\1\2\3"'
+
+        add :open_quote,
+            :pattern => /(^|\(|\s)("|\\")(\S+)/iue,
+            :replacement => "\\1#{QUOTE_FIRS_OPEN}\\3"
+
+        add :close_quote,
+            :pattern => /([a-zа-я0-9]|\.|&hellip;|!|\?)("|\\")+(\.|&hellip;|;|:|\?|!|,|\s|\)|<\/|$)/uie,
+            :replacement => "\\1#{QUOTE_FIRS_CLOSE}\\3"
+      end
+
+    end
+  end
+end

data/lib/typograph/string_sub_each.rb

@@ -0,0 +1,20 @@
+require "strscan"
+
+class String
+  def sub_each regexp, &block
+    return unless block_given?
+    scanner = StringScanner.new self
+    while scanner.scan(regexp)
+      pre_match = scanner.pre_match
+      replacement = block[scanner]
+      new_pos = pre_match.length + replacement.length
+      scanner.string = pre_match << replacement << scanner.post_match
+      scanner.pos = new_pos
+    end
+    scanner.string
+  end
+
+  def sub_each! regexp, &block
+    self.replace(self.sub_each(regexp, &block))
+  end
+end

data/test/typograph_test.rb

@@ -0,0 +1,90 @@
+$:.unshift File.join(File.dirname(__FILE__), '..', 'lib')
+
+require "test/unit"
+require "typograph"
+
+class TypographTest < Test::Unit::TestCase
+
+  include Typograph
+  include Typograph::Rules
+
+  def test_dash_rules
+    assert_equal "как-то", DashRules.run("как-то")
+    assert_equal "привет&nbsp;&mdash; привет", DashRules.run("привет - привет")
+    assert_equal "с 2000&ndash;3000", DashRules.run("с 2000-3000")
+    assert_equal "изза", DashRules.run("изза")
+    assert_equal " из-за ", DashRules.run(" из -за ")
+    assert_equal " как-то ", DashRules.run(" как- то ")
+  end
+
+  def test_etc_rules
+    assert_equal "anjlab&trade;", EtcRules.run("anjlab(tm)")
+    assert_equal "anjlab&copy;&nbsp;", EtcRules.run("anjlab(c) ")
+    assert_equal "приве&#769;т", EtcRules.run("приве`т")
+  end
+
+  def test_math_rules
+    assert_equal "2&times;5", MathRules.run("2x5")
+    assert_equal " &frac12; ", MathRules.run(" 1/2 ")
+    assert_equal " &frac14; ", MathRules.run(" 1/4 ")
+    assert_equal " &frac34; ", MathRules.run(" 3/4 ")
+    assert_equal " &frac34;&#8800;&frac12; ", MathRules.run(" 3/4!=1/2 ")
+    assert_equal " &frac34; + &frac14; + &frac12; ", MathRules.run(" 3/4 + 1/4 + 1/2 ")
+    assert_equal "&plusmn;", MathRules.run("+-")
+    assert_equal "&plusmn;5", MathRules.run("+-5")
+  end
+
+  # punchamark rules tests
+  def test_auto_comma
+    assert_equal "вы будете, а я нет", PunchmarkRules.run("вы будете а я нет")
+    assert_equal "вы будете, а я нет", PunchmarkRules.run("вы будете, а я нет")
+  end
+
+  def test_punctuation_marks_limit
+    assert_equal "hello!!!", PunchmarkRules.run("hello!!!!!!!!!!!!!!!!!!")
+    assert_equal "WTF???", PunchmarkRules.run("WTF???????")
+  end
+
+  def test_punctuation_marks_base_limit
+    assert_equal ":", PunchmarkRules.run(":::")
+    assert_equal ",", PunchmarkRules.run(",,")
+  end
+
+  def test_hellip
+    assert_equal "&hellip;", PunchmarkRules.run("..")
+    assert_equal "&hellip;", PunchmarkRules.run("...")
+    assert_equal "&hellip;", PunchmarkRules.run("....")
+  end
+
+  def test_eng_apostrophe
+    assert_equal "she&rsquo;s", PunchmarkRules.run("she's")
+  end
+
+  def test_fix_brackets
+    assert_equal "(круто)", PunchmarkRules.run("( круто )")
+  end
+
+  def test_quote_rules
+    assert_equal "&laquo;как-то&raquo;", QuoteRules.run('"как-то"')
+    assert_equal "&laquo;как-то&raquo; &laquo;как-то&raquo;", QuoteRules.run('"как-то" "как-то"')
+    assert_equal "привет всем &laquo;крутым&raquo; штукам", QuoteRules.run('привет всем "крутым" штукам')
+    assert_equal "&laquo;как-то &laquo;как-то&raquo;", QuoteRules.run('"как-то "как-то""')
+  end
+
+  def test_complex_processing
+    src = '"Кока-Кола" впервые появилась в СССР в 1979 году в
+    ходе подготовки Олимпийских игр в Москве. Её автор -
+    фармацевт Джон Стит Пембертон, бывший офицер американской
+    Армии конфедерации!!!! ( сам напиток был придуман в 1886 году
+    в штате Джорджия ) и дальше какая-то байда....'
+
+    dest = '&laquo;Кока-Кола&raquo; впервые появилась в СССР в 1979 году в
+    ходе подготовки Олимпийских игр в Москве. Её автор&nbsp;&mdash;
+    фармацевт Джон Стит Пембертон, бывший офицер американской
+    Армии конфедерации!!! (сам напиток был придуман в 1886 году
+    в штате Джорджия) и дальше какая-то байда&hellip;'
+
+    assert_equal dest, Typograph.run(src)
+  end
+
+end

data/typograph.gemspec

@@ -0,0 +1,52 @@
+# Generated by jeweler
+# DO NOT EDIT THIS FILE
+# Instead, edit Jeweler::Tasks in Rakefile, and run `rake gemspec`
+# -*- encoding: utf-8 -*-
+
+Gem::Specification.new do |s|
+  s.name = %q{typograph}
+  s.version = "0.1.1"
+
+  s.required_rubygems_version = Gem::Requirement.new(">= 0") if s.respond_to? :required_rubygems_version=
+  s.authors = ["Yury Korolev"]
+  s.date = %q{2009-08-25}
+  s.description = %q{apply some basic typography rules to text}
+  s.email = %q{yury.korolev@gmail.com}
+  s.extra_rdoc_files = [
+    "README"
+  ]
+  s.files = [
+    ".gitignore",
+     "README",
+     "Rakefile",
+     "VERSION",
+     "lib/typograph.rb",
+     "lib/typograph/rules.rb",
+     "lib/typograph/rules/dash_rules.rb",
+     "lib/typograph/rules/etc_rules.rb",
+     "lib/typograph/rules/math_rules.rb",
+     "lib/typograph/rules/punchmark_rules.rb",
+     "lib/typograph/rules/quote_rules.rb",
+     "lib/typograph/string_sub_each.rb",
+     "test/typograph_test.rb",
+     "typograph.gemspec"
+  ]
+  s.homepage = %q{http://github.com/yury/typograph}
+  s.rdoc_options = ["--charset=UTF-8"]
+  s.require_paths = ["lib"]
+  s.rubygems_version = %q{1.3.5}
+  s.summary = %q{ruby html typograph}
+  s.test_files = [
+    "test/typograph_test.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

metadata

@@ -0,0 +1,67 @@
+--- !ruby/object:Gem::Specification 
+name: yury-typograph
+version: !ruby/object:Gem::Version 
+  version: 0.1.1
+platform: ruby
+authors: 
+- Yury Korolev
+autorequire: 
+bindir: bin
+cert_chain: []
+
+date: 2009-08-25 00:00:00 -07:00
+default_executable: 
+dependencies: []
+
+description: apply some basic typography rules to text
+email: yury.korolev@gmail.com
+executables: []
+
+extensions: []
+
+extra_rdoc_files: 
+- README
+files: 
+- .gitignore
+- README
+- Rakefile
+- VERSION
+- lib/typograph.rb
+- lib/typograph/rules.rb
+- lib/typograph/rules/dash_rules.rb
+- lib/typograph/rules/etc_rules.rb
+- lib/typograph/rules/math_rules.rb
+- lib/typograph/rules/punchmark_rules.rb
+- lib/typograph/rules/quote_rules.rb
+- lib/typograph/string_sub_each.rb
+- test/typograph_test.rb
+- typograph.gemspec
+has_rdoc: false
+homepage: http://github.com/yury/typograph
+licenses: 
+post_install_message: 
+rdoc_options: 
+- --charset=UTF-8
+require_paths: 
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement 
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      version: "0"
+  version: 
+required_rubygems_version: !ruby/object:Gem::Requirement 
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      version: "0"
+  version: 
+requirements: []
+
+rubyforge_project: 
+rubygems_version: 1.3.5
+signing_key: 
+specification_version: 3
+summary: ruby html typograph
+test_files: 
+- test/typograph_test.rb