virastar 0.0.3 → 0.0.4

data/TODO

@@ -1,12 +1,10 @@
-- do not put space after : in the context of numbers like hour 19:45 => ۱۹:۴۵
-- spacing after , : ; causing a lot of problem in this case (,) => (, )
+- destories /n if the line ends with ""
 - do not destroy urls dots and colons
+/https?://([-\w\.]+)+(:\d+)?(/([\w/_\.]*(\?\S+)?)?)?/
 
 - translate to js
 
 
 DONE:
-- he yeh => hamzeh should consider arabic yeh and zwnj chars
-- replacing quotes shouldn't be greedy 
-- for suffixes also consider tari
-- (IMP)  do not replace \n probably caused by fixing spaces after dots and commas
+- spacing after , : ; causing a lot of problem in this case (,) => (, )
+- do not put space after : in the context of numbers like hour 19:45 => ۱۹:۴۵

data/lib/virastar.rb

@@ -47,15 +47,6 @@ module Virastar
       # remove unnecessary zwnj char that are succeeded/preceded by a space
       text.gsub!(/\s+‌|‌\s+/,' ') if @cleanup_zwnj
 
-      # should fix spacing for () [] {}  “” «»
-      if @fix_spacing_for_braces_and_quotes
-        text.gsub!(/\s*(\()\s*([^)]+?)\s*?(\))\s*/,' \1\2\3 ')
-        text.gsub!(/\s*(\[)\s*([^)]+?)\s*?(\])\s*/,' \1\2\3 ')
-        text.gsub!(/\s*(\{)\s*([^)]+?)\s*?(\})\s*/,' \1\2\3 ')
-        text.gsub!(/\s*(“)\s*([^)]+?)\s*?(”)\s*/,' \1\2\3 ')
-        text.gsub!(/\s*(«)\s*([^)]+?)\s*?(»)\s*/,' \1\2\3 ')
-      end
-
       # character replacement
       persian_numbers = "۱۲۳۴۵۶۷۸۹۰"
       arabic_numbers  = "١٢٣٤٥٦٧٨٩٠"
@@ -101,12 +92,25 @@ module Virastar
       # : ; , . ! ? and their persian equivalents should have one space after and no space before
       if @fix_spacing_for_braces_and_quotes
         text.gsub!(/[ ‌	]*([:;,؛،.؟!]{1})[ ‌	]*/, '\1 ')
+        # do not put space after colon that separates time parts
+        text.gsub!(/([۰-۹]+):\s+([۰-۹]+)/, '\1:\2')
+      end
+      
+      
+
+      # should fix spacing for () [] {}  “” «»
+      if @fix_spacing_for_braces_and_quotes
+        text.gsub!(/\s*(\()\s*([^)]+?)\s*?(\))\s*/,' \1\2\3 ')
+        text.gsub!(/\s*(\[)\s*([^)]+?)\s*?(\])\s*/,' \1\2\3 ')
+        text.gsub!(/\s*(\{)\s*([^)]+?)\s*?(\})\s*/,' \1\2\3 ')
+        text.gsub!(/\s*(“)\s*([^)]+?)\s*?(”)\s*/,' \1\2\3 ')
+        text.gsub!(/\s*(«)\s*([^)]+?)\s*?(»)\s*/,' \1\2\3 ')
       end
 
       # should replace more than one space with just a single one
       if @cleanup_spacing
         text.gsub!(/[ ]+/,' ')
-        #text.gsub!(/\s*[\n]+\s*/," \n")
+        text.gsub!(/([\n]+)[ 	‌]*/,'\1')
       end
 
       # remove spaces, tabs, and new lines from the beginning and enf of file

data/lib/virastar/version.rb

@@ -1,3 +1,3 @@
 module Virastar
-  VERSION = "0.0.3"
+  VERSION = "0.0.4"
 end

data/spec/virastar_spec.rb

@@ -145,19 +145,19 @@ describe Virastar do
   end
 
   it "should replace more that one line breaks with just one" do
-    test  = "this is \n \n \n     \n a test"
-    result = "this is \na test"
-    test2  = "this is\n\n\n\na test"
-    result2 = "this is \na test"
-    test3  = "this is \n\n\n\n    a test"
-    result3 = "this is \na test"
+    test    = "this is \n \n \n     \n a test"
+    result  = "this is \n\n\n\na test"
+    test2   = "this is\n\n\n\na test"
+    result2 = "this is\n\n\n\na test"
+    test3   = "this is \n\n\n    a test"
+    result3 = "this is \n\n\na test"
 
     test.persian_cleanup.should  == result
     test2.persian_cleanup.should  == result2
     test3.persian_cleanup.should  == result3
   end
 
-  it "should not replace line breaks" do
+  it "should not replace line breaks and should remove spaces after line break" do
     test  = "this is \n  a test"
     result = "this is \na test"
     test.persian_cleanup.should  == result
@@ -181,6 +181,30 @@ describe Virastar do
     test.persian_cleanup.should  == result
   end
 
+  it "should not create spacing for something like (,)" do
+    test = "this is (,) comma"
+    result = "this is (،) comma"
+    test.persian_cleanup.should  == result
+  end
+
+  it "should not puts space after time colon separator" do
+    test = "12:34"
+    result = "۱۲:۳۴"
+    test.persian_cleanup.should  == result
+  end
+
+  it "should not destroy URLs"
+   # do
+   #    test = "http://virastar.heroku.com"
+   #    result = "http://virastar.heroku.com"
+   #    test.persian_cleanup.should  == result
+   #end
+
+  it "should not replace line breaks when the line ends with quotes" do
+    test = 'استفاده از "گيومه های فارسي"\nساده است'
+    result = 'استفاده از «گیومه‌های فارسی» \nساده است'
+    test.persian_cleanup.should  == result
+  end
 
   context "aggressive editing" do
     it "should replace more than one ! or ? mark with just one" do

metadata

@@ -1,13 +1,13 @@
 --- !ruby/object:Gem::Specification 
 name: virastar
 version: !ruby/object:Gem::Version 
-  hash: 25
+  hash: 23
   prerelease: 
   segments: 
   - 0
   - 0
-  - 3
-  version: 0.0.3
+  - 4
+  version: 0.0.4
 platform: ruby
 authors: 
 - Allen A. Bargi
@@ -15,7 +15,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2011-01-20 00:00:00 +01:00
+date: 2011-01-22 00:00:00 +01:00
 default_executable: 
 dependencies: 
 - !ruby/object:Gem::Dependency