textutils 0.9.0 → 0.9.1

data/Manifest.txt

@@ -4,6 +4,8 @@ README.md
 Rakefile
 lib/textutils.rb
 lib/textutils/classifier.rb
+lib/textutils/core_ext/file.rb
+lib/textutils/core_ext/time.rb
 lib/textutils/filter/code_filter.rb
 lib/textutils/filter/comment_filter.rb
 lib/textutils/filter/erb_django_filter.rb

data/lib/textutils.rb

@@ -42,6 +42,9 @@ require 'textutils/helper/address_helper'
 require 'textutils/helper/value_helper'
 
 require 'textutils/utils'
+require 'textutils/core_ext/file'
+require 'textutils/core_ext/time'
+
 require 'textutils/reader/code_reader'
 require 'textutils/reader/hash_reader'
 require 'textutils/reader/hash_reader_v2'

data/lib/textutils/core_ext/file.rb

@@ -0,0 +1,15 @@
+# encoding: utf-8
+
+class File
+  def self.read_utf8( path )
+    text = open( path, 'r:bom|utf-8' ) do |file|
+      file.read
+    end
+
+    # NB: for convenience: convert fancy unicode dashes/hyphens to plain ascii hyphen-minus
+    text = TextUtils.convert_unicode_dashes_to_plain_ascii( text, path: path )
+
+    text
+  end
+end # class File
+

data/lib/textutils/core_ext/time.rb

@@ -0,0 +1,19 @@
+# encoding: utf-8
+
+
+class Time
+  
+  def self.cet( str )   # central european time (cet) + central european summer time (cest)  
+    ActiveSupport::TimeZone['Vienna'].parse( str )
+  end
+
+  def self.eet( str )  # eastern european time (eet)  + 2 hours
+    ActiveSupport::TimeZone['Bucharest'].parse( str )
+  end
+  
+  def self.cst( str )  # central standard time (cst) - 6 hours 
+    ActiveSupport::TimeZone['Mexico City'].parse( str )
+  end
+  
+end # class Time
+

data/lib/textutils/helper/title_helper.rb

@@ -45,7 +45,7 @@ module TextUtils
 
   def strip_whitespaces( title )
       # remove all whitespace and punctuation
-      title.gsub( /[ \t_\-\.!()\[\]'"\/]/, '' )
+      title.gsub( /[ \t_\-\.!()\[\]'"’\/]/, '' )
   end
 
   def strip_special_chars( title )
@@ -81,21 +81,28 @@ module TextUtils
       ## see http://en.wikipedia.org/wiki/List_of_XML_and_HTML_character_entity_references  for more
       
       ## todo: add unicode codepoint name
-      
+
+
       alternatives = [
         ['ß', 'ss'],
+
         ['æ', 'ae'],
         ['ä', 'ae'],
-        ['ā', 'a' ],  # e.g. Liepājas
+        ['ā', 'a' ],  # e.g. Liepājas, Kāṭhmāḍaũ
         ['á', 'a' ],  # e.g. Bogotá, Králové
+        ['à', 'a' ],  # e.g. Thành Phố Hồ Chí Minh [Saigon]
         ['ã', 'a' ],  # e.g  São Paulo
         ['ă', 'a' ],  # e.g. Chișinău
         ['â', 'a' ],  # e.g  Goiânia
         ['å', 'a' ],  # e.g. Vålerenga
         ['ą', 'a' ],  # e.g. Śląsk
+
         ['ç', 'c' ],  # e.g. São Gonçalo, Iguaçu, Neftçi
         ['ć', 'c' ],  # e.g. Budućnost
         ['č', 'c' ],  # e.g. Tradiční, Výčepní
+
+        ['ḍ', 'd' ],  # e.g. Kāṭhmāḍaũ [Kathmandu]
+
         ['é', 'e' ],  # e.g. Vélez, Králové
         ['è', 'e' ],  # e.g. Rivières
         ['ê', 'e' ],  # e.g. Grêmio
@@ -103,31 +110,53 @@ module TextUtils
         ['ĕ', 'e' ],  # e.g. Svĕtlý
         ['ė', 'e' ],  # e.g. Vėtra
         ['ë', 'e' ],  # e.g. Skënderbeu
+
         ['ğ', 'g' ],  # e.g. Qarabağ
+
+        ['ḥ', 'h' ],  # e.g. Ad-Dawḥah [Doha]
+
         ['ì', 'i' ],  # e.g. Potosì
         ['í', 'i' ],  # e.g. Ústí
         ['ï', 'i' ],  # e.g. El Djazaïr
+        ['ī', 'i' ],  # e.g. Al-Iskandarīyah [Alexandria]
+
         ['ł', 'l' ],  # e.g. Wisła, Wrocław
         ['ñ', 'n' ],  # e.g. Porteño
         ['ň', 'n' ],  # e.g. Plzeň, Třeboň
+
         ['ö', 'oe'],
         ['ő', 'o' ],  # e.g. Győri
         ['ó', 'o' ],  # e.g. Colón, Łódź, Kraków
         ['õ', 'o' ],  # e.g. Nõmme
         ['ô', 'o' ],  # e.g. Amazônia (pt)
-        ['ō', 'o' ],  # e.g. Tōkyō
+        ['ō', 'o' ],  # e.g. Tōkyō, Pishōr
+        ['ŏ', 'o' ],  # e.g. P'yŏngyang [Pyongyang]
         ['ø', 'o' ],  # e.g. Fuglafjørdur, København
+        ['ố', 'o' ],  # e.g. Thành Phố Hồ Chí Minh [Saigon]
+        ['ồ', 'o' ],  # e.g. Thành Phố Hồ Chí Minh [Saigon]
+        ['ộ', 'o' ],  # e.g. Hà Nội [Hanoi]
+
         ['ř', 'r' ],  # e.g. Třeboň
+
         ['ș', 's' ],  # e.g. Chișinău, București
         ['ş', 's' ],  # e.g. Beşiktaş
         ['š', 's' ],  # e.g. Košice
+        ['ṣ', 's' ], # e.g. Al-Mawṣil [Mosul]
+
         ['ť', 't' ],  # e.g. Měšťan
+        ['ṭ', 't' ],  # e.g. Al-Kharṭūm [Khartoum], Kāṭhmāḍaũ
+
         ['ü', 'ue'],
         ['ú', 'u' ],  # e.g. Fútbol
         ['ù', 'u' ],  # e.g. Xyauyù (it)
         ['ū', 'u' ],  # e.g. Sūduva
         ['ů', 'u' ],  # e.g. Sládkův
+        ['ũ', 'u' ],  # e.g. Kāṭhmāḍaũ [Kathmandu]
+
         ['ı', 'u' ],  # e.g. Bakı   # use u?? (Baku) why-why not?
+
+        ['x̌', 'x'],   # e.g. Pex̌awar  [Peshawar]
+
         ['ý', 'y' ],  # e.g. Nefitrovaný
         ['ź', 'z' ],  # e.g. Łódź
         ['ž', 'z' ],  # e.g. Domžale, Petržalka
@@ -135,15 +164,25 @@ module TextUtils
 
         ['Á', 'a' ],  # e.g. Águila (es)
         ['Č', 'c' ],  # e.g. České
+
+        ['Ḥ', 'h' ],  # e.g. Ḥalab [Aleppo]
+        ['Ḫ', 'h' ],  # e.g. Ḫamīs Mušayṭ
         ['İ', 'i' ],  # e.g. İnter
         ['Í', 'i' ],  # e.g. ÍBV
         ['Ł', 'l' ],  # e.g. Łódź
+
         ['Ö', 'oe' ], # e.g. Örebro
+        ['Ō', 'o' ],  # e.g. Ōsaka [Osaka]
         ['Ø', 'o' ],  # e.g. Nogne Ø Imperial Stout (no)
+
         ['Ř', 'r' ],  # e.g. Řezák
+
         ['Ś', 's' ],  # e.g. Śląsk
         ['Š', 's' ],  # e.g. MŠK
         ['Ş', 's' ],  # e.g. Şüvälan
+        ['Ṣ', 's' ],  # e.g. Ṣan'ā' [Sana'a]
+
+        ['Ṭ', 't' ],  # e.g. Ṭarābulus [Tripoli]
         ['Ú', 'u' ],  # e.g. Ústí, Újpest
         ['Ž', 'z' ],   # e.g. Žilina
         ['Ż', 'z' ]    # e.g. Żywiec (polish)

data/lib/textutils/utils.rb

@@ -9,21 +9,6 @@ end
 
 
 
-class File
-  def self.read_utf8( path )
-    text = open( path, 'r:bom|utf-8' ) do |file|
-      file.read
-    end
-
-    # NB: for convenience: convert fancy unicode dashes/hyphens to plain ascii hyphen-minus
-    text = TextUtils.convert_unicode_dashes_to_plain_ascii( text, path: path )
-
-    text
-  end
-end # class File
-
-
-
 def title_esc_regex( title_unescaped )
   puts "*** warn: depreceated fn call: use TextUtils.title_esc_regex() or include TextUtils::TitleHelpers"
   TextUtils.title_esc_regex( title_unescaped )

data/lib/textutils/version.rb

@@ -1,6 +1,6 @@
 
 module TextUtils
 
-  VERSION = '0.9.0'
+  VERSION = '0.9.1'
 
 end   # module TextUtils

data/test/test_title_helper.rb

@@ -38,7 +38,37 @@ class TestTitleHelper < MiniTest::Unit::TestCase
       [ '$Alianz$ Arena', 'alianzarena' ],
       [ 'Arena Amazônia', 'arenaamazonia' ],
       [ 'Tōkyō [Tokyo]', 'tokyo' ],
-      [ 'El Djazaïr [Algiers]', 'eldjazair' ]
+      [ 'Ōsaka [Osaka]', 'osaka' ],
+      [ 'El Djazaïr [Algiers]', 'eldjazair' ],
+      [ 'Al-Kharṭūm [Khartoum]', 'alkhartum' ],
+      [ 'Ṭarābulus [Tripoli]', 'tarabulus' ],
+      [ 'Al-Iskandarīyah [Alexandria]', 'aliskandariyah' ],
+      [ 'Pex̌awar', 'pexawar'],
+      [ 'Pishōr', 'pishor' ],
+      [ 'Pishāwar', 'pishawar' ],
+      [ 'Islām ābād', 'islamabad' ],
+      [ 'Thành Phố Hồ Chí Minh [Saigon]', 'thanhphohochiminh' ],
+      [ 'Hà Nội [Hanoi]', 'hanoi' ],
+      [ 'Donets’k', 'donetsk' ],
+      [ 'Baghdād [Baghdad]', 'baghdad'],
+      [ 'Al-Mawṣil [Mosul]', 'almawsil'],
+      [ 'Al-Baṣrah [Basra]', 'albasrah'],
+      [ 'Arbīl [Erbil]', 'arbil' ],
+      [ 'Kirkūk [Kirkuk]', 'kirkuk' ],
+      [ 'Tehrān [Tehran]', 'tehran' ],
+      [ 'Eṣfahān [Isfahan]', 'esfahan' ],
+      [ 'Shīrāz [Shiraz]', 'shiraz' ],
+      [ 'Tabrīz [Tabriz]', 'tabriz' ],
+      [ 'Ahvāz [Ahvaz]', 'ahvaz' ],
+      [ 'Ad-Dawḥah [Doha]', 'addawhah'],
+      [ 'Ḥalab [Aleppo]', 'halab'],
+      [ 'Al-Madīnah [Medina]', 'almadinah'],
+      [ 'Ad-Dammām [Dammam]', 'addammam' ],
+      [ 'Aṭ-Ṭā’if', 'attaif'], 
+      [ 'Ḫamīs Mušayṭ', 'hamismusayt'],
+      [ "Ṣan'ā' [Sana'a]", 'sana'],
+      [ "P'yŏngyang [Pyongyang]", 'pyongyang' ],
+      [ 'Kāṭhmāḍaũ [Kathmandu]', 'kathmadau' ]
     ]
 
     txt_io.each do |txt|

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: textutils
 version: !ruby/object:Gem::Version
-  version: 0.9.0
+  version: 0.9.1
   prerelease: 
 platform: ruby
 authors:
@@ -13,7 +13,7 @@ date: 2014-03-24 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: props
-  requirement: &76193510 !ruby/object:Gem::Requirement
+  requirement: &73970010 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -21,10 +21,10 @@ dependencies:
         version: '0'
   type: :runtime
   prerelease: false
-  version_requirements: *76193510
+  version_requirements: *73970010
 - !ruby/object:Gem::Dependency
   name: logutils
-  requirement: &76193110 !ruby/object:Gem::Requirement
+  requirement: &73969260 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
@@ -32,10 +32,10 @@ dependencies:
         version: '0.5'
   type: :runtime
   prerelease: false
-  version_requirements: *76193110
+  version_requirements: *73969260
 - !ruby/object:Gem::Dependency
   name: activesupport
-  requirement: &76192860 !ruby/object:Gem::Requirement
+  requirement: &73969070 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -43,10 +43,10 @@ dependencies:
         version: '0'
   type: :runtime
   prerelease: false
-  version_requirements: *76192860
+  version_requirements: *73969070
 - !ruby/object:Gem::Dependency
   name: rdoc
-  requirement: &76192430 !ruby/object:Gem::Requirement
+  requirement: &73968810 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
@@ -54,10 +54,10 @@ dependencies:
         version: '3.10'
   type: :development
   prerelease: false
-  version_requirements: *76192430
+  version_requirements: *73968810
 - !ruby/object:Gem::Dependency
   name: hoe
-  requirement: &76191590 !ruby/object:Gem::Requirement
+  requirement: &73953840 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
@@ -65,7 +65,7 @@ dependencies:
         version: '3.3'
   type: :development
   prerelease: false
-  version_requirements: *76191590
+  version_requirements: *73953840
 description: textutils - Text Filters, Helpers, Readers and More
 email: ruby-talk@ruby-lang.org
 executables: []
@@ -79,6 +79,8 @@ files:
 - Rakefile
 - lib/textutils.rb
 - lib/textutils/classifier.rb
+- lib/textutils/core_ext/file.rb
+- lib/textutils/core_ext/time.rb
 - lib/textutils/filter/code_filter.rb
 - lib/textutils/filter/comment_filter.rb
 - lib/textutils/filter/erb_django_filter.rb