amatch 0.2.3 → 0.2.4

data/lib/amatch/version.rb

@@ -0,0 +1,8 @@
+module Amatch
+  # Amatch version
+  VERSION         = '0.2.4'
+  VERSION_ARRAY   = VERSION.split(/\./).map { |x| x.to_i } # :nodoc:
+  VERSION_MAJOR   = VERSION_ARRAY[0] # :nodoc:
+  VERSION_MINOR   = VERSION_ARRAY[1] # :nodoc:
+  VERSION_BUILD   = VERSION_ARRAY[2] # :nodoc:
+end

data/tests/test_hamming.rb

@@ -1,7 +1,7 @@
 require 'test/unit'
 require 'amatch'
 
-class TC_Hamming < Test::Unit::TestCase
+class TestHamming < Test::Unit::TestCase
   include Amatch
 
   D = 0.000001
@@ -56,4 +56,3 @@ class TC_Hamming < Test::Unit::TestCase
     assert_in_delta 1.0, @long.similar(@long.pattern), D
   end
 end
-  # vim: set et sw=2 ts=2:

data/tests/test_jaro.rb

@@ -0,0 +1,29 @@
+require 'test/unit'
+require File.dirname(__FILE__) + "/../ext/amatch"
+
+class TestJaro < Test::Unit::TestCase
+  include Amatch
+
+  D = 0.0005
+
+  def setup
+    @martha = Jaro.new('Martha')
+    @dwayne = Jaro.new('dwayne')
+    @dixon  = Jaro.new('DIXON')
+    @one    = Jaro.new('one')
+  end
+
+  def test_case
+    @martha.ignore_case = true
+    assert_in_delta 0.944, @martha.match('MARHTA'), D
+    @martha.ignore_case = false
+    assert_in_delta 0.444, @martha.match('MARHTA'), D
+  end
+
+  def test_match
+    assert_in_delta 0.944, @martha.match('MARHTA'), D
+    assert_in_delta 0.822, @dwayne.match('DUANE'), D
+    assert_in_delta 0.767, @dixon.match('DICKSONX'), D
+    assert_in_delta 0.667, @one.match('orange'), D
+  end
+end

data/tests/test_jaro_winkler.rb

@@ -0,0 +1,38 @@
+require 'test/unit'
+require 'amatch'
+
+class TestJaroWinkler < Test::Unit::TestCase
+  include Amatch
+
+  D = 0.0005
+
+  def setup
+    @martha = JaroWinkler.new('Martha')
+    @dwayne = JaroWinkler.new('dwayne')
+    @dixon  = JaroWinkler.new('DIXON')
+    @one    = JaroWinkler.new("one")
+  end
+
+  def test_case
+    @martha.ignore_case = true
+    assert_in_delta 0.961, @martha.match('MARHTA'), D
+    @martha.ignore_case = false
+    assert_in_delta 0.500, @martha.match('MARHTA'), D
+  end
+
+  def test_match
+    assert_in_delta 0.961, @martha.match('MARHTA'), D
+    assert_in_delta 0.840, @dwayne.match('DUANE'), D
+    assert_in_delta 0.813, @dixon.match('DICKSONX'), D
+    assert_in_delta 0, @one.match('two'), D
+    assert_in_delta 0.700, @one.match('orange'), D
+  end
+
+  def test_scaling_factor
+    assert_in_delta 0.1, @martha.scaling_factor, 0.0000001
+    @martha.scaling_factor = 0.2
+    assert_in_delta 0.978, @martha.match('MARHTA'), D
+    @martha.scaling_factor = 0.5 # this is far too high
+    assert_in_delta 1.028, @martha.match('MARHTA'), D
+  end
+end

data/tests/test_levenshtein.rb

@@ -1,23 +1,16 @@
 require 'test/unit'
 require 'amatch'
 
-class TC_Levenshtein < Test::Unit::TestCase
+class TestLevenshtein < Test::Unit::TestCase
   include Amatch
 
-  D = 0.000001
-  
   def setup
+    @d        = 0.000001
     @empty    = Levenshtein.new('')
     @simple   = Levenshtein.new('test')
     @long     = Levenshtein.new('A' * 160)
   end
 
-  def test_long
-    a = "lost this fantasy,  this fantasy,  this  fantasy,  this fantasy,  this fantasy,  this fantasy\r\n\r\nGood love  Neat work\r\n\r\nSuper job Fancy work\r\n\r\nPants job Cool work"
-    b = "lost\r\n\r\nGood love Neat work\r\n\r\nSuper  job Fancy work\r\n\r\nPants job Cool work"
-    p a.levenshtein_similar b 
-  end
-
   def test_match
     assert_equal 4,     @simple.match('')
     assert_equal 0,     @simple.match('test')
@@ -60,26 +53,31 @@ class TC_Levenshtein < Test::Unit::TestCase
   end
 
   def test_similar
-    assert_in_delta 1, @empty.similar(''), D
-    assert_in_delta 0, @empty.similar('not empty'), D
-    assert_in_delta 0.0, @simple.similar(''), D
-    assert_in_delta 1.0, @simple.similar('test'), D
-    assert_in_delta 0.8, @simple.similar('testa'), D
-    assert_in_delta 0.8, @simple.similar('atest'), D
-    assert_in_delta 0.8, @simple.similar('teast'), D
-    assert_in_delta 0.75, @simple.similar('est'), D
-    assert_in_delta 0.75, @simple.similar('tes'), D
-    assert_in_delta 0.75, @simple.similar('tst'), D
-    assert_in_delta 0.75, @simple.similar('best'), D
-    assert_in_delta 0.75, @simple.similar('tost'), D
-    assert_in_delta 0.75, @simple.similar('tesa'), D
-    assert_in_delta 0.25, @simple.similar('taex'), D
-    assert_in_delta 0.4, @simple.similar('aaatestbbb'), D
-    assert_in_delta 0.75, @simple.pattern.levenshtein_similar('est'), D
+    assert_in_delta 1, @empty.similar(''), @d
+    assert_in_delta 0, @empty.similar('not empty'), @d
+    assert_in_delta 0.0, @simple.similar(''), @d
+    assert_in_delta 1.0, @simple.similar('test'), @d
+    assert_in_delta 0.8, @simple.similar('testa'), @d
+    assert_in_delta 0.8, @simple.similar('atest'), @d
+    assert_in_delta 0.8, @simple.similar('teast'), @d
+    assert_in_delta 0.75, @simple.similar('est'), @d
+    assert_in_delta 0.75, @simple.similar('tes'), @d
+    assert_in_delta 0.75, @simple.similar('tst'), @d
+    assert_in_delta 0.75, @simple.similar('best'), @d
+    assert_in_delta 0.75, @simple.similar('tost'), @d
+    assert_in_delta 0.75, @simple.similar('tesa'), @d
+    assert_in_delta 0.25, @simple.similar('taex'), @d
+    assert_in_delta 0.4, @simple.similar('aaatestbbb'), @d
+    assert_in_delta 0.75, @simple.pattern.levenshtein_similar('est'), @d
   end
 
   def test_long
-    assert_in_delta 1.0, @long.similar(@long.pattern), D
+    assert_in_delta 1.0, @long.similar(@long.pattern), @d
+  end
+
+  def test_long2
+    a = "lost this fantasy,  this fantasy,  this  fantasy,  this fantasy,  this fantasy,  this fantasy\r\n\r\nGood love  Neat work\r\n\r\nSuper job Fancy work\r\n\r\nPants job Cool work"
+    b = "lost\r\n\r\nGood love Neat work\r\n\r\nSuper  job Fancy work\r\n\r\nPants job Cool work"
+    assert a.levenshtein_similar(b)
   end
 end
-  # vim: set et sw=2 ts=2:

data/tests/test_longest_subsequence.rb

@@ -1,7 +1,7 @@
 require 'test/unit'
 require 'amatch'
 
-class TC_LongestSubsequence < Test::Unit::TestCase
+class TestLongestSubsequence < Test::Unit::TestCase
   include Amatch
 
   D = 0.000001
@@ -59,4 +59,3 @@ class TC_LongestSubsequence < Test::Unit::TestCase
     assert_in_delta 1.0, @long.similar(@long.pattern), D
   end
 end
-  # vim: set et sw=2 ts=2:

data/tests/test_longest_substring.rb

@@ -1,7 +1,7 @@
 require 'test/unit'
 require 'amatch'
 
-class TC_LongestSubstring < Test::Unit::TestCase
+class TestLongestSubstring < Test::Unit::TestCase
   include Amatch
 
   D = 0.000001
@@ -59,4 +59,3 @@ class TC_LongestSubstring < Test::Unit::TestCase
     assert_in_delta 1.0, @long.similar(@long.pattern), D
   end
 end
-  # vim: set et sw=2 ts=2:

data/tests/test_pair_distance.rb

@@ -1,7 +1,7 @@
 require 'test/unit'
 require 'amatch'
 
-class TC_PairDistance < Test::Unit::TestCase
+class TestPairDistance < Test::Unit::TestCase
   include Amatch
 
   D = 0.000001
@@ -84,4 +84,3 @@ class TC_PairDistance < Test::Unit::TestCase
     assert_in_delta 1.0, @long.similar(@long.pattern), D
   end
 end
-  # vim: set et sw=2 ts=2:

data/tests/test_sellers.rb

@@ -2,49 +2,48 @@ require 'test/unit'
 require 'amatch'
 require 'test_levenshtein'
 
-class TC_Sellers < TC_Levenshtein
+class TestSellers < TestLevenshtein
   include Amatch
-
-  D = 0.000001
-
+  
   def setup
+    @d        = 0.000001
     @empty    = Sellers.new('')
     @simple   = Sellers.new('test')
     @long     = Sellers.new('A' * 160)
   end
 
   def test_weights
-    assert_in_delta 1, @simple.substitution, D
-    assert_in_delta 1, @simple.insertion, D
-    assert_in_delta 1, @simple.deletion, D
+    assert_in_delta 1, @simple.substitution, @d
+    assert_in_delta 1, @simple.insertion, @d
+    assert_in_delta 1, @simple.deletion, @d
     @simple.insertion = 1
     @simple.substitution = @simple.deletion = 1000
-    assert_in_delta 1, @simple.match('tst'), D
-    assert_in_delta 1, @simple.search('bbbtstccc'), D
+    assert_in_delta 1, @simple.match('tst'), @d
+    assert_in_delta 1, @simple.search('bbbtstccc'), @d
     @simple.deletion = 1
     @simple.substitution = @simple.insertion = 1000
-    assert_in_delta 1, @simple.match('tedst'), D
-    assert_in_delta 1, @simple.search('bbbtedstccc'), D
+    assert_in_delta 1, @simple.match('tedst'), @d
+    assert_in_delta 1, @simple.search('bbbtedstccc'), @d
     @simple.substitution = 1
     @simple.deletion = @simple.insertion = 1000
-    assert_in_delta 1, @simple.match('tast'), D
-    assert_in_delta 1, @simple.search('bbbtastccc'), D
+    assert_in_delta 1, @simple.match('tast'), @d
+    assert_in_delta 1, @simple.search('bbbtastccc'), @d
     @simple.insertion = 0.5
     @simple.substitution = @simple.deletion = 1000
-    assert_in_delta 0.5, @simple.match('tst'), D
-    assert_in_delta 0.5, @simple.search('bbbtstccc'), D
+    assert_in_delta 0.5, @simple.match('tst'), @d
+    assert_in_delta 0.5, @simple.search('bbbtstccc'), @d
     @simple.deletion = 0.5
     @simple.substitution = @simple.insertion = 1000
-    assert_in_delta 0.5, @simple.match('tedst'), D
-    assert_in_delta 0.5, @simple.search('bbbtedstccc'), D
+    assert_in_delta 0.5, @simple.match('tedst'), @d
+    assert_in_delta 0.5, @simple.search('bbbtedstccc'), @d
     @simple.substitution = 0.5
     @simple.deletion = @simple.insertion = 1000
-    assert_in_delta 0.5, @simple.match('tast'), D
-    assert_in_delta 0.5, @simple.search('bbbtastccc'), D
+    assert_in_delta 0.5, @simple.match('tast'), @d
+    assert_in_delta 0.5, @simple.search('bbbtastccc'), @d
     @simple.reset_weights
-    assert_in_delta 1, @simple.substitution, D
-    assert_in_delta 1, @simple.insertion, D
-    assert_in_delta 1, @simple.deletion, D
+    assert_in_delta 1, @simple.substitution, @d
+    assert_in_delta 1, @simple.insertion, @d
+    assert_in_delta 1, @simple.deletion, @d
   end
 
   def test_weight_exceptions
@@ -54,45 +53,44 @@ class TC_Sellers < TC_Levenshtein
   end
 
   def test_similar
-    assert_in_delta 0.0, @simple.similar(''), D
-    assert_in_delta 1.0, @simple.similar('test'), D
-    assert_in_delta 0.8, @simple.similar('testa'), D
-    assert_in_delta 0.8, @simple.similar('atest'), D
-    assert_in_delta 0.8, @simple.similar('teast'), D
-    assert_in_delta 0.75, @simple.similar('est'), D
-    assert_in_delta 0.75, @simple.similar('tes'), D
-    assert_in_delta 0.75, @simple.similar('tst'), D
-    assert_in_delta 0.75, @simple.similar('best'), D
-    assert_in_delta 0.75, @simple.similar('tost'), D
-    assert_in_delta 0.75, @simple.similar('tesa'), D
-    assert_in_delta 0.25, @simple.similar('taex'), D
-    assert_in_delta 0.4, @simple.similar('aaatestbbb'), D
-    assert_in_delta 0.75, @simple.pattern.levenshtein_similar('est'), D
+    assert_in_delta 0.0, @simple.similar(''), @d
+    assert_in_delta 1.0, @simple.similar('test'), @d
+    assert_in_delta 0.8, @simple.similar('testa'), @d
+    assert_in_delta 0.8, @simple.similar('atest'), @d
+    assert_in_delta 0.8, @simple.similar('teast'), @d
+    assert_in_delta 0.75, @simple.similar('est'), @d
+    assert_in_delta 0.75, @simple.similar('tes'), @d
+    assert_in_delta 0.75, @simple.similar('tst'), @d
+    assert_in_delta 0.75, @simple.similar('best'), @d
+    assert_in_delta 0.75, @simple.similar('tost'), @d
+    assert_in_delta 0.75, @simple.similar('tesa'), @d
+    assert_in_delta 0.25, @simple.similar('taex'), @d
+    assert_in_delta 0.4, @simple.similar('aaatestbbb'), @d
+    assert_in_delta 0.75, @simple.pattern.levenshtein_similar('est'), @d
   end
 
   def test_similar
-    assert_in_delta 1, @empty.similar(''), D
-    assert_in_delta 0, @empty.similar('not empty'), D
-    assert_in_delta 0.0, @simple.similar(''), D
-    assert_in_delta 1.0, @simple.similar('test'), D
-    assert_in_delta 0.8, @simple.similar('testa'), D
-    assert_in_delta 0.8, @simple.similar('atest'), D
-    assert_in_delta 0.8, @simple.similar('teast'), D
-    assert_in_delta 0.75, @simple.similar('est'), D
-    assert_in_delta 0.75, @simple.similar('tes'), D
-    assert_in_delta 0.75, @simple.similar('tst'), D
-    assert_in_delta 0.75, @simple.similar('best'), D
-    assert_in_delta 0.75, @simple.similar('tost'), D
-    assert_in_delta 0.75, @simple.similar('tesa'), D
-    assert_in_delta 0.25, @simple.similar('taex'), D
-    assert_in_delta 0.4, @simple.similar('aaatestbbb'), D
+    assert_in_delta 1, @empty.similar(''), @d
+    assert_in_delta 0, @empty.similar('not empty'), @d
+    assert_in_delta 0.0, @simple.similar(''), @d
+    assert_in_delta 1.0, @simple.similar('test'), @d
+    assert_in_delta 0.8, @simple.similar('testa'), @d
+    assert_in_delta 0.8, @simple.similar('atest'), @d
+    assert_in_delta 0.8, @simple.similar('teast'), @d
+    assert_in_delta 0.75, @simple.similar('est'), @d
+    assert_in_delta 0.75, @simple.similar('tes'), @d
+    assert_in_delta 0.75, @simple.similar('tst'), @d
+    assert_in_delta 0.75, @simple.similar('best'), @d
+    assert_in_delta 0.75, @simple.similar('tost'), @d
+    assert_in_delta 0.75, @simple.similar('tesa'), @d
+    assert_in_delta 0.25, @simple.similar('taex'), @d
+    assert_in_delta 0.4, @simple.similar('aaatestbbb'), @d
     @simple.insertion = 1
     @simple.substitution = @simple.deletion = 2
-    assert_in_delta 0.875, @simple.similar('tst'), D
+    assert_in_delta 0.875, @simple.similar('tst'), @d
   end
 
   def test_long
-    assert_in_delta 1.0, @long.similar(@long.pattern), D
+    assert_in_delta 1.0, @long.similar(@long.pattern), @d
   end
 end
-  # vim: set et sw=2 ts=2:

metadata

@@ -1,71 +1,92 @@
 --- !ruby/object:Gem::Specification 
-rubygems_version: 0.8.11
-specification_version: 1
 name: amatch
 version: !ruby/object:Gem::Version 
-  version: 0.2.3
-date: 2006-09-26 00:00:00 +02:00
-summary: Approximate String Matching library
-require_paths: 
-- ext
-email: flori@ping.de
-homepage: http://amatch.rubyforge.org
-rubyforge_project: amatch
-description: Amatch is a library for approximate string matching and searching in strings. Several algorithms can be used to do this, and it's also possible to compute a similarity metric number between 0.0 and 1.0 for two given strings.
-autorequire: amatch
-default_executable: agrep.rb
-bindir: bin
-has_rdoc: true
-required_ruby_version: !ruby/object:Gem::Version::Requirement 
-  requirements: 
-  - - ">"
-    - !ruby/object:Gem::Version 
-      version: 0.0.0
-  version: 
+  version: 0.2.4
 platform: ruby
-signing_key: 
-cert_chain: 
 authors: 
 - Florian Frank
+autorequire: 
+bindir: bin
+cert_chain: []
+
+date: 2009-08-26 00:00:00 +02:00
+default_executable: agrep.rb
+dependencies: []
+
+description: |
+  Amatch is a library for approximate string matching and searching in strings.
+  Several algorithms can be used to do this, and it's also possible to compute a
+  similarity metric number between 0.0 and 1.0 for two given strings.
+
+email: flori@ping.de
+executables: 
+- agrep.rb
+extensions: 
+- ext/extconf.rb
+extra_rdoc_files: 
+- ext/amatch.c
+- lib/amatch/version.rb
+- doc-main.txt
 files: 
-- bin
-- VERSION
-- tests
-- GPL
-- README.en
-- install.rb
-- ext
-- Rakefile
 - CHANGES
 - bin/agrep.rb
-- tests/runner.rb
+- VERSION
+- README
+- ext/common.h
+- ext/extconf.rb
+- ext/amatch.c
+- ext/pair.h
+- ext/pair.c
+- Rakefile
+- lib/amatch/version.rb
 - tests/test_longest_substring.rb
-- tests/test_levenshtein.rb
 - tests/test_hamming.rb
-- tests/test_sellers.rb
-- tests/test_pair_distance.rb
 - tests/test_longest_subsequence.rb
-- ext/tags
-- ext/pair.h
-- ext/extconf.rb
-- ext/MANIFEST
-- ext/pair.c
-- ext/amatch.c
-test_files: 
-- tests/runner.rb
+- tests/test_pair_distance.rb
+- tests/test_levenshtein.rb
+- tests/test_jaro.rb
+- tests/test_sellers.rb
+- tests/test_jaro_winkler.rb
+- COPYING
+- install.rb
+- doc-main.txt
+has_rdoc: true
+homepage: http://amatch.rubyforge.org
+licenses: []
+
+post_install_message: 
 rdoc_options: 
-- --title
-- Amatch -- Approximate Matching
 - --main
-- Amatch
-- --line-numbers
-extra_rdoc_files: []
-
-executables: 
-- agrep.rb
-extensions: 
-- ext/extconf.rb
+- doc-main.txt
+- --title
+- amatch - Approximate Matching
+require_paths: 
+- ext
+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: []
 
-dependencies: []
-
+rubyforge_project: amatch
+rubygems_version: 1.3.4
+signing_key: 
+specification_version: 3
+summary: Approximate String Matching library
+test_files: 
+- tests/test_longest_substring.rb
+- tests/test_hamming.rb
+- tests/test_longest_subsequence.rb
+- tests/test_pair_distance.rb
+- tests/test_levenshtein.rb
+- tests/test_jaro.rb
+- tests/test_sellers.rb
+- tests/test_jaro_winkler.rb