ActionItem70: transliterating UTF-8 strings

git-svn-id: https://svn.colivre.coop.br/svn/noosfero/trunk@518 3f533792-8f58-4932-b0fe-aaf55b0a4547

ActionItem70: transliterating UTF-8 strings
git-svn-id: https://svn.colivre.coop.br/svn/noosfero/trunk@518 3f533792-8f58-4932-b0fe-aaf55b0a4547
AntonioTerceiro
1 parent ba3df928
Showing 2 changed files with 50 additions and 0 deletions Show diff stats
lib/noosfero/transliterations.rb
test/unit/transliterations_test.rb
@@ -0,0 +1,39 @@
+module Noosfero::Transliterations
+
+  TRANSLATION = {
+    [ 'Á', 'À', 'À', 'Â', 'Ã', 'Ä' ] => 'A',
+    [ 'á', 'à', 'à', 'â', 'ã', 'ä', 'ª' ] => 'a',
+    [ 'É', 'È', 'Ê', 'Ë' ] => 'E',
+    [ 'é', 'è', 'ê', 'ë' ] => 'e',
+    [ 'Í', 'Ì', 'Î', 'Ï' ] => 'I',
+    [ 'í', 'ì', 'î', 'ï' ] => 'i',
+    [ 'Ó', 'Ò', 'Ô', 'Ö', 'Õ', 'º' ] => 'O',
+    [ 'ó', 'ò', 'ô', 'ö', 'õ', 'º' ] => 'o',
+    [ 'Ú', 'Ù', 'Û', 'Ü' ] => 'U',
+    [ 'ú', 'ù', 'û', 'ü' ] => 'u',
+    [ 'Ç' ] => 'C',
+    [ 'ç' ] => 'c',
+    [ 'Ñ' ] => 'N',
+    [ 'ñ' ] => 'n',
+    [ 'Ÿ' ] => 'Y',
+    [ 'ÿ' ] => 'y',
+  }
+
+  # transliterate a string (assumed to be contain UTF-8 data)
+  # into ASCII by replacing non-ascii characters to their
+  # ASCII.
+  #
+  # The transliteration is, of course, lossy, and its performance is poor.
+  def transliterate
+
+    new = self.clone
+    Noosfero::Transliterations::TRANSLATION.each { |from,to|
+      from.each { |seq|
+        new.gsub!(seq, to)
+      }
+    }
+    new
+  end
+end
+
+String.send(:include, Noosfero::Transliterations)
@@ -0,0 +1,11 @@
+require File.dirname(__FILE__) + '/../test_helper'
+require 'noosfero/transliterations'
+
+class TransliterationsTest < Test::Unit::TestCase
+
+  def test_should_transliterate
+    assert_equal 'eeeeEEOOoocaaaiIIiuuyYnN', 'éèëêÊËÖÔöôçäàâîÏÎïûüÿŸñÑ'.transliterate
+  end
+
+
+end
...	...	@@ -0,0 +1,39 @@
	1	+module Noosfero::Transliterations
	2	+
	3	+ TRANSLATION = {
	4	+ [ 'Á', 'À', 'À', 'Â', 'Ã', 'Ä' ] => 'A',
	5	+ [ 'á', 'à', 'à', 'â', 'ã', 'ä', 'ª' ] => 'a',
	6	+ [ 'É', 'È', 'Ê', 'Ë' ] => 'E',
	7	+ [ 'é', 'è', 'ê', 'ë' ] => 'e',
	8	+ [ 'Í', 'Ì', 'Î', 'Ï' ] => 'I',
	9	+ [ 'í', 'ì', 'î', 'ï' ] => 'i',
	10	+ [ 'Ó', 'Ò', 'Ô', 'Ö', 'Õ', 'º' ] => 'O',
	11	+ [ 'ó', 'ò', 'ô', 'ö', 'õ', 'º' ] => 'o',
	12	+ [ 'Ú', 'Ù', 'Û', 'Ü' ] => 'U',
	13	+ [ 'ú', 'ù', 'û', 'ü' ] => 'u',
	14	+ [ 'Ç' ] => 'C',
	15	+ [ 'ç' ] => 'c',
	16	+ [ 'Ñ' ] => 'N',
	17	+ [ 'ñ' ] => 'n',
	18	+ [ 'Ÿ' ] => 'Y',
	19	+ [ 'ÿ' ] => 'y',
	20	+ }
	21	+
	22	+ # transliterate a string (assumed to be contain UTF-8 data)
	23	+ # into ASCII by replacing non-ascii characters to their
	24	+ # ASCII.
	25	+ #
	26	+ # The transliteration is, of course, lossy, and its performance is poor.
	27	+ def transliterate
	28	+
	29	+ new = self.clone
	30	+ Noosfero::Transliterations::TRANSLATION.each { \|from,to\|
	31	+ from.each { \|seq\|
	32	+ new.gsub!(seq, to)
	33	+ }
	34	+ }
	35	+ new
	36	+ end
	37	+end
	38	+
	39	+String.send(:include, Noosfero::Transliterations)
...	...
...	...	@@ -0,0 +1,11 @@
	1	+require File.dirname(__FILE__) + '/../test_helper'
	2	+require 'noosfero/transliterations'
	3	+
	4	+class TransliterationsTest < Test::Unit::TestCase
	5	+
	6	+ def test_should_transliterate
	7	+ assert_equal 'eeeeEEOOoocaaaiIIiuuyYnN', 'éèëêÊËÖÔöôçäàâîÏÎïûüÿŸñÑ'.transliterate
	8	+ end
	9	+
	10	+
	11	+end
...	...