kevintyll-ofac 1.1.10 → 1.1.11

data/History.txt

@@ -58,9 +58,15 @@
     in the ofac_sdn database.  This method ignores address and city and does not produce a score.
     Usage: Ofac.new({:name => 'Oscar Hernandez', :city => 'Clearwater', :address => '123 somewhere ln'}).db_hit?
 
-
 == 1.1.10 2009-07-28
 
 * 1 minor enhancement:
   * Modified the select in OfacSdn to use the city to pull records from the database.  This is another compromise to improve performance,
-    but still get a score.  The db_hit? method is still faster, but this will still calculate a score for accuracy.
+    but still get a score.  The db_hit? method is still faster, but this will still calculate a score for accuracy.
+
+== 1.1.11 2009-07-29
+
+* 1 minor enhancement:
+  * Removed the changes from 1.1.10.  1.1.10 produced a score of 0 when a higher score was more accurate.
+    Changed the scoring algorithm to not include partial matches on sounds like.  This code has little impact on score, and
+    has a very high overhead on performance.

data/VERSION.yml

@@ -1,4 +1,4 @@
 --- 
 :minor: 1
-:patch: 10
+:patch: 11
 :major: 1

data/lib/ofac/models/ofac.rb

@@ -147,20 +147,12 @@ class Ofac
         sql_alt_name_partial = name_array.collect {|partial_name| ["alternate_identity_name like ?", "%#{partial_name}%"]}
         conditions = sql_name_partial + sql_alt_name_partial
         conditions = conditions.transpose
-        unless @identity[:city].blank?
-          city_condition = ' and (city like ? or city is null)'
-        end
-        name_values = conditions.second
-        conditions = conditions.first.join(' or ')
-        conditions = "(#{conditions})"
-        conditions += city_condition if city_condition
-        condition_array = [conditions] + name_values
-        condition_array += ["%#{@identity[:city]}%"] if city_condition
-        
-        possible_sdns = OfacSdn.find_all_by_sdn_type('individual',:select => 'name, alternate_identity_name, address, city', :conditions => condition_array)
+        conditions = [conditions.first.join(' or ')] + conditions.second
+
+        possible_sdns = OfacSdn.find_all_by_sdn_type('individual',:select => 'name, alternate_identity_name, address, city', :conditions => conditions)
         possible_sdns = possible_sdns.collect {|sdn|{:name => "#{sdn['name']}|#{sdn['alternate_identity_name']}", :city => sdn['city'], :address => sdn['address']}}
      
-        match = OfacMatch.new({:name => {:weight => 60, :token => "#{name_array.join(' ')}"},
+        match = OfacMatch.new({:name => {:weight => 60, :token => "#{name_array.join(', ')}"},
             :address => {:weight => 10, :token => @identity[:address]},
             :city => {:weight => 30, :token => @identity[:city]}})
 
@@ -175,10 +167,10 @@ class Ofac
   def process_name
     #you can pass in a full name, or specify the first and last name
     if @identity[:name].kind_of?(Hash)
-      name_array = [@identity[:name][:first_name],@identity[:name][:last_name]].compact
+      name_array = [@identity[:name][:last_name],@identity[:name][:first_name]].compact
     else
       partial_name = @identity[:name].gsub(/\W/,'|')
-      name_array = partial_name.split('|')
+      name_array = partial_name.split('|').reverse
     end
   end
 

data/lib/ofac/ofac_match.rb

@@ -11,7 +11,7 @@ class OfacMatch
   #        :zip    => {:weight => 10, :token => '33759', :type => :number}})
   #
   # data hash keys:
-  #   * <tt>data[:weight]</tt> - value to apply to the score if there is a match (Default is 100/number of key in the record hash)
+  #   * <tt>data[:weight]</tt> - value to apply to the score if there is a match (Default is 100/number of keys in the record hash)
   #   * <tt>data[:token]</tt> - string to match
   #   * <tt>data[:match]</tt> - set from records hash
   #   * <tt>data[:score]</tt> - output field
@@ -79,12 +79,20 @@ class OfacMatch
               value = data[:token] == data[:match] ? 1 : 0
             else
               #first see if there is an exact match
-              value = data[:token] == data[:match] ? 1 : 0
+              value = data[:match].split('|').include?(data[:token]) ? 1 : 0
 
               unless value > 0
                 #do a sounds like with the data as given to see if we get a match
                 #if match on sounds_like, only give .75 of the weight.
-                value = data[:token].ofac_sounds_like(data[:match],false) ? 0.75 : 0
+                data[:match].split('|').each do |separate_value|
+                  if data[:token].ofac_sounds_like(separate_value,false)
+                    value = 0.75
+                    break
+                  else
+                    value = 0
+                  end
+                end
+
               end
                 
               #if no match, then break the data down and see if we can find matches on the
@@ -105,16 +113,6 @@ class OfacMatch
                   #first see if we get an exact match of the partial
                   if success = match_array.include?(partial_token)
                     value += partial_weight
-                  else
-                    #otherwise, see if the partial sounds like any part of the OFAC record
-                    match_array.each do |partial_match|
-                      if partial_match.ofac_sounds_like(partial_token,false)
-                        #give partial value for every part of token that is matched.
-                        value += partial_weight * 0.75
-                        success = true
-                        break
-                      end
-                    end
                   end
                   unless success
                     #if this for :address or :city

data/test/ofac_test.rb

@@ -73,20 +73,6 @@ class OfacTest < Test::Unit::TestCase
       assert_equal 90, Ofac.new({:name => 'Oscar Hernandez', :city => 'Clearwater'}).score
     end
 
-    should "not find a match if the city does not match" do
-      assert_equal 0, Ofac.new({:name => 'Oscar Hernandez', :city => 'Tampa'}).score
-      #only name matches
-      assert_equal 0, Ofac.new({:name => 'Oscar Hernandez', :city => 'no match', :address => 'no match'}).score
-    end
-
-    should "find a match if the city is not passed in" do
-      assert_equal 60, Ofac.new({:name => 'Oscar Hernandez'}).score
-    end
-
-    should "find a match if the city is not in the database" do
-      assert_equal 60, Ofac.new({:name => 'Raul AGUIAR'}).score
-    end
-
     should "give a score of 100 if there is a name and city and address match" do
       assert_equal 100, Ofac.new({:name => 'Oscar Hernandez', :city => 'Clearwater', :address => '123 somewhere ln'}).score
     end
@@ -96,16 +82,15 @@ class OfacTest < Test::Unit::TestCase
       #32456 summer lane sounds like 32456 Somewhere ln so is adds 75% of the address weight to the score, or 8.
       assert_equal 98, Ofac.new({:name => 'Oscar Hernandez', :city => 'Clearwater', :address => '32456 summer lane'}).score
 
-      #summer sounds like somewhere, and all numbers sound alike, so 2 of the 3 address elements match by sound.
-      #Each element is worth 10\3 or 3.33.  Exact matches add 2.33 each, and the sounds like adds 2.33 * .75 or 2.5
-      #because sounds like matches only add 75% of it's weight.
-      #2.5 + 2.5 = 5
-      assert_equal 95, Ofac.new({:name => 'Oscar Hernandez', :city => 'Clearwater', :address => '12358 summer blvd'}).score
-
+      #Louis Eduardo Lopez Mendez sounds like Luis Eduardo Lopez Mendez:
+      #:name has a weight of 60, so a sounds like is worth 45
+      assert_equal 45, Ofac.new({:name => {:first_name => 'Louis Eduardo', :last_name => 'Lopez Mendez'}, :city => 'Las Vegas', :address => 'no match'}).score
+    end
 
-      #Louis sounds like Luis, and Lopez is an exact match:
-      #:name has a weight of 60, so each element is worth 30.  A sounds like match is worth 30 * .75
-      assert_equal 53, Ofac.new({:name => 'Louis Lopez', :city => 'Las Vegas', :address => 'no match'}).score
+    should 'not give partial scores if sounds like does not match the entire string' do
+      #summer sounds like somewhere, and all numbers sound alike, so 2 of the 3 address elements match by sound
+      #but the whole thing does not make a sounds like match so only city matches, subtract 1 for no address match
+      assert_equal 89, Ofac.new({:name => 'Oscar Hernandez', :city => 'Clearwater', :address => '12358 summer blvd'}).score
     end
 
     should "return an array of possible hits" do

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification 
 name: kevintyll-ofac
 version: !ruby/object:Gem::Version 
-  version: 1.1.10
+  version: 1.1.11
 platform: ruby
 authors: 
 - Kevin Tyll
@@ -9,7 +9,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2009-07-28 00:00:00 -07:00
+date: 2009-07-29 00:00:00 -07:00
 default_executable: 
 dependencies: []