picky 4.12.2 → 4.12.3

data/lib/picky/character_substituters/base.rb

@@ -0,0 +1,21 @@
+# -*- coding: utf-8 -*-
+
+module Picky
+
+  module CharacterSubstituters
+
+    class Base
+
+      def initialize
+        @chars = ActiveSupport::Multibyte.proxy_class
+      end
+
+      def to_s
+        self.class.name
+      end
+
+    end
+
+  end
+
+end

data/lib/picky/character_substituters/polish.rb

@@ -0,0 +1,24 @@
+# -*- coding: utf-8 -*-
+
+module Picky
+
+  module CharacterSubstituters
+
+    class Polish < Base
+
+      def substitute text
+        trans = @chars.new(text).normalize :kd
+
+        trans.gsub! 'Ł', 'L'
+        trans.gsub! 'ł', 'l'
+
+        trans.unpack('U*').select { |cp|
+          cp < 0x0300 || cp > 0x035F
+        }.pack 'U*'
+      end
+
+    end
+
+  end
+
+end

data/lib/picky/character_substituters/west_european.rb

@@ -10,11 +10,7 @@ module Picky
     # ä, ö, ü => ae, oe, ue.
     # (and more, see specs)
     #
-    class WestEuropean
-
-      def initialize
-        @chars = ActiveSupport::Multibyte.proxy_class
-      end
+    class WestEuropean < Base
 
       # Substitutes occurrences of certain characters
       # (like Umlauts) with ASCII representations of them.
@@ -45,10 +41,6 @@ module Picky
         }.pack 'U*'
       end
 
-      def to_s
-        self.class.name
-      end
-
     end
 
   end

data/lib/picky/loader.rb

@@ -268,7 +268,9 @@ module Picky
         load_relative 'tokenizer/regexp_wrapper'
         load_relative 'tokenizer'
         # load_relative 'rack/harakiri' # Needs to be explicitly loaded/required.
+        load_relative 'character_substituters/base'
         load_relative 'character_substituters/west_european'
+        load_relative 'character_substituters/polish'
         load_relative 'splitters/automatic'
         load_generators
         load_inner_api

data/spec/lib/character_substituters/polish_spec.rb

@@ -0,0 +1,74 @@
+# encoding: utf-8
+#
+require 'spec_helper'
+
+describe Picky::CharacterSubstituters::Polish do
+
+  let(:substituter) { described_class.new.tap { |s| s.substitute '' } }
+
+  # A bit of metaprogramming to help with the myriads of its.
+  #
+  def self.it_should_substitute special_character, normal_character
+    it "should substitute #{special_character} with #{normal_character}" do
+      substituter.substitute(special_character).should == normal_character
+    end
+  end
+  def self.it_should_not_substitute special_character
+    it "should not substitute #{special_character}" do
+      substituter.substitute(special_character).should == special_character
+    end
+  end
+
+  # Speed spec at the top since the order of the describes made the
+  # speed spec trip. And not on mushrooms either.
+  #
+  describe "speed" do
+    it "is fast" do
+      substituter.substitute 'ą' # Prerun
+      result = performance_of { substituter.substitute('ą') }
+      result.should < 0.00009
+    end
+    it "is fast" do
+      result = performance_of { substituter.substitute('abcdefghijklmnopqrstuvwxyz1234567890') }
+      result.should < 0.00015
+    end
+  end
+
+  describe 'to_s' do
+    it 'outputs correctly' do
+      substituter.to_s.should == 'Picky::CharacterSubstituters::Polish'
+    end
+  end
+
+  describe "normal characters" do
+    it_should_not_substitute('abcdefghijklmnopqrstuvwxyz1234567890')
+  end
+
+  describe "situations" do
+    it_should_substitute 'Michał Prawda', 'Michal Prawda'
+    it_should_substitute 'Brzęczyszczykiewicz', 'Brzeczyszczykiewicz'
+  end
+
+  describe "diacritics" do
+    #ĄąĘęĆ棳ŃńŚśÓ󏟯ż
+    it_should_substitute 'ą', 'a'
+    it_should_substitute 'Ą', 'A'
+    it_should_substitute 'ę', 'e'
+    it_should_substitute 'Ę', 'E'
+    it_should_substitute 'ć', 'c'
+    it_should_substitute 'Ć', 'C'
+    it_should_substitute 'ł', 'l'
+    it_should_substitute 'Ł', 'L'
+    it_should_substitute 'ń', 'n'
+    it_should_substitute 'Ń', 'N'
+    it_should_substitute 'ś', 's'
+    it_should_substitute 'Ś', 'S'
+    it_should_substitute 'ó', 'o'
+    it_should_substitute 'Ó', 'O'
+    it_should_substitute 'ź', 'z'
+    it_should_substitute 'Ź', 'Z'
+    it_should_substitute 'ż', 'z'
+    it_should_substitute 'Ż', 'Z'
+  end
+
+end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: picky
 version: !ruby/object:Gem::Version
-  version: 4.12.2
+  version: 4.12.3
   prerelease: 
 platform: ruby
 authors:
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2012-12-03 00:00:00.000000000 Z
+date: 2012-12-06 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rspec
@@ -34,7 +34,7 @@ dependencies:
     requirements:
     - - ~>
       - !ruby/object:Gem::Version
-        version: 4.12.2
+        version: 4.12.3
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
@@ -42,7 +42,7 @@ dependencies:
     requirements:
     - - ~>
       - !ruby/object:Gem::Version
-        version: 4.12.2
+        version: 4.12.3
 - !ruby/object:Gem::Dependency
   name: text
   requirement: !ruby/object:Gem::Requirement
@@ -163,6 +163,8 @@ files:
 - lib/picky/category_indexed.rb
 - lib/picky/category_indexing.rb
 - lib/picky/category_realtime.rb
+- lib/picky/character_substituters/base.rb
+- lib/picky/character_substituters/polish.rb
 - lib/picky/character_substituters/west_european.rb
 - lib/picky/console.rb
 - lib/picky/constants.rb
@@ -331,6 +333,7 @@ files:
 - spec/lib/category_indexing_spec.rb
 - spec/lib/category_realtime_spec.rb
 - spec/lib/category_spec.rb
+- spec/lib/character_substituters/polish_spec.rb
 - spec/lib/character_substituters/west_european_spec.rb
 - spec/lib/extensions/array_spec.rb
 - spec/lib/extensions/hash_spec.rb
@@ -496,6 +499,7 @@ test_files:
 - spec/lib/category_indexing_spec.rb
 - spec/lib/category_realtime_spec.rb
 - spec/lib/category_spec.rb
+- spec/lib/character_substituters/polish_spec.rb
 - spec/lib/character_substituters/west_european_spec.rb
 - spec/lib/extensions/array_spec.rb
 - spec/lib/extensions/hash_spec.rb