loose_tight_dictionary 0.2.2 → 0.2.3

data/lib/loose_tight_dictionary/cached_result.rb

@@ -0,0 +1,74 @@
+class LooseTightDictionary
+  class CachedResult < ::ActiveRecord::Base
+    set_table_name :loose_tight_dictionary_cached_results
+    
+    def self.create_table
+      connection.create_table :loose_tight_dictionary_cached_results do |t|
+        t.string :a_class
+        t.string :a
+        t.string :b_class
+        t.string :b
+      end
+      connection.add_index :loose_tight_dictionary_cached_results, [:a_class, :b_class, :a], :name => 'aba'
+      connection.add_index :loose_tight_dictionary_cached_results, [:a_class, :b_class, :b], :name => 'abb'
+      connection.add_index :loose_tight_dictionary_cached_results, [:a_class, :b_class, :a, :b], :name => 'abab'
+      reset_column_information
+    end
+    
+    def self.setup(from_scratch = false)
+      connection.drop_table :loose_tight_dictionary_cached_results if from_scratch and table_exists?
+      create_table unless table_exists?
+    end
+    
+    module ActiveRecordBaseExtension
+      # required options:
+      # :primary_key - what to call on this class
+      # :foreign_key - what to call on the other class
+      def cache_loose_tight_dictionary_matches_with(other_active_record_class, options)
+        other = other_active_record_class.to_s.singularize.camelcase
+        me = name
+        if me < other
+          a = me
+          b = other
+          primary_key = :a
+          foreign_key = :b
+        else
+          a = other
+          b = me
+          primary_key = :b
+          foreign_key = :a
+        end
+
+        # def aircraft
+        define_method other.underscore.pluralize do
+          other.constantize.where options[:foreign_key] => send("#{other.underscore.pluralize}_foreign_keys")
+        end
+  
+        # def flight_segments_foreign_keys
+        define_method "#{other.underscore.pluralize}_foreign_keys" do
+          fz = ::LooseTightDictionary::CachedResult.arel_table
+          sql = fz.project(fz[foreign_key]).where(fz["#{primary_key}_class".to_sym].eq(self.class.name).and(fz["#{foreign_key}_class".to_sym].eq(other)).and(fz[primary_key].eq(send(options[:primary_key])))).to_sql
+          connection.select_values sql
+        end
+  
+        # def cache_aircraft!
+        define_method "cache_#{other.underscore.pluralize}!" do
+          other_class = other.constantize
+          primary_key_value = send options[:primary_key]
+          other_class.loose_tight_dictionary.find_all(primary_key_value).each do |other_instance|
+            attrs = {}
+            attrs[primary_key] = primary_key_value
+            attrs["#{primary_key}_class"] = self.class.name
+            attrs[foreign_key] = other_instance.send options[:foreign_key]
+            attrs["#{foreign_key}_class"] = other
+            unless ::LooseTightDictionary::CachedResult.exists? attrs
+              ::LooseTightDictionary::CachedResult.create! attrs
+            end
+          end
+        end
+      end
+    end
+  end
+end
+
+::ActiveRecord::Base.extend ::LooseTightDictionary::CachedResult::ActiveRecordBaseExtension

data/lib/loose_tight_dictionary/version.rb

@@ -1,3 +1,3 @@
 class LooseTightDictionary
-  VERSION = '0.2.2'
+  VERSION = '0.2.3'
 end

data/lib/loose_tight_dictionary.rb

@@ -18,6 +18,7 @@ class LooseTightDictionary
   autoload :Wrapper, 'loose_tight_dictionary/wrapper'
   autoload :Similarity, 'loose_tight_dictionary/similarity'
   autoload :Score, 'loose_tight_dictionary/score'
+  autoload :CachedResult, 'loose_tight_dictionary/cached_result'
   
   class Freed < RuntimeError; end
   

data/loose_tight_dictionary.gemspec

@@ -21,7 +21,11 @@ Gem::Specification.new do |s|
 
   s.add_development_dependency "shoulda"
   s.add_development_dependency "remote_table"
-  s.add_dependency 'activesupport', '>=2.3.4'
+  s.add_development_dependency 'activerecord', '>=3'
+  s.add_development_dependency 'mysql'
+  s.add_development_dependency 'cohort_scope'
+  s.add_development_dependency 'weighted_average'
+  s.add_dependency 'activesupport', '>=3'
   s.add_dependency 'amatch'
   s.add_dependency 'to_regexp', '>=0.0.3'
 end

data/test/test_cache.rb

@@ -0,0 +1,130 @@
+require 'helper'
+
+require 'active_support/all'
+require 'active_record'
+require 'cohort_scope'
+require 'weighted_average'
+
+ActiveRecord::Base.establish_connection(
+  'adapter' => 'mysql',
+  'database' => 'loose_tight_dictionary_test',
+  'username' => 'root',
+  'password' => 'password'
+)
+
+# ActiveRecord::Base.logger = Logger.new $stderr
+
+ActiveSupport::Inflector.inflections do |inflect|
+  inflect.uncountable 'aircraft'
+end
+
+require 'loose_tight_dictionary/cached_result'
+
+::LooseTightDictionary::CachedResult.setup(true)
+::LooseTightDictionary::CachedResult.delete_all
+
+class Aircraft < ActiveRecord::Base
+  set_primary_key :icao_code
+  
+  cache_loose_tight_dictionary_matches_with :flight_segments, :primary_key => :aircraft_description, :foreign_key => :aircraft_description
+    
+  def aircraft_description
+    [manufacturer_name, model_name].compact.join(' ')
+  end
+  
+  def self.loose_tight_dictionary
+    @loose_tight_dictionary ||= LooseTightDictionary.new all, :haystack_reader => lambda { |straw| straw.aircraft_description }
+  end
+  
+  def self.create_table
+    connection.drop_table(:aircraft) rescue nil
+    connection.execute %{
+CREATE TABLE `aircraft` (
+  `icao_code` varchar(255) DEFAULT NULL,
+  `manufacturer_name` varchar(255) DEFAULT NULL,
+  `model_name` varchar(255) DEFAULT NULL,
+  PRIMARY KEY (`icao_code`)
+) ENGINE=InnoDB DEFAULT CHARSET=utf8;
+    }
+    reset_column_information
+  end
+end
+
+class FlightSegment < ActiveRecord::Base
+  set_primary_key :row_hash
+  
+  cache_loose_tight_dictionary_matches_with :aircraft, :primary_key => :aircraft_description, :foreign_key => :aircraft_description
+  
+  extend CohortScope
+  self.minimum_cohort_size = 1
+  
+  def self.create_table
+    connection.drop_table(:flight_segments) rescue nil
+    connection.execute %{
+CREATE TABLE `flight_segments` (
+  `row_hash` varchar(255) NOT NULL DEFAULT '',
+  `aircraft_description` varchar(255) DEFAULT NULL,
+  `passengers` int(11) DEFAULT NULL,
+  `seats` int(11) DEFAULT NULL,
+  PRIMARY KEY (`row_hash`)
+) ENGINE=InnoDB DEFAULT CHARSET=utf8;
+    }
+  end
+end
+
+FlightSegment.create_table
+Aircraft.create_table
+
+a = Aircraft.new
+a.icao_code = 'B742'
+a.manufacturer_name = 'Boeing'
+a.model_name = '747-200'
+a.save!
+
+fs = FlightSegment.new
+fs.row_hash = 'madison to chicago'
+fs.aircraft_description = 'BORING 747200'
+fs.passengers = 10
+fs.seats = 10
+fs.save!
+
+fs = FlightSegment.new
+fs.row_hash = 'madison to minneapolis'
+fs.aircraft_description = 'bing 747'
+fs.passengers = 100
+fs.seats = 5
+fs.save!
+
+FlightSegment.find_each do |fs|
+  fs.cache_aircraft!
+end
+
+class TestCache < Test::Unit::TestCase
+  def test_002_one_degree_of_separation
+    aircraft = Aircraft.find('B742')
+    assert_equal 2, aircraft.flight_segments.count
+  end
+  
+  def test_003_standard_sql_calculations
+    aircraft = Aircraft.find('B742')
+    assert_equal 110, aircraft.flight_segments.sum(:passengers)
+  end
+  
+  def test_004_weighted_average
+    aircraft = Aircraft.find('B742')
+    assert_equal 5.4545, aircraft.flight_segments.weighted_average(:seats, :weighted_by => :passengers)
+  end
+  
+  def test_005_right_way_to_do_cohorts
+    aircraft = Aircraft.find('B742')
+    assert_equal 2, FlightSegment.big_cohort(:aircraft_description => aircraft.flight_segments_foreign_keys).count
+  end
+  
+  def test_006_you_can_get_aircraft_from_flight_segments
+    fs = FlightSegment.first
+    # you need to add an aircraft_description column
+    assert_raises(ActiveRecord::StatementInvalid) do
+      assert_equal 2, fs.aircraft.count
+    end
+  end
+end

metadata

@@ -2,7 +2,7 @@
 name: loose_tight_dictionary
 version: !ruby/object:Gem::Version 
   prerelease: 
-  version: 0.2.2
+  version: 0.2.3
 platform: ruby
 authors: 
 - Seamus Abshere
@@ -10,7 +10,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2011-05-02 00:00:00 -05:00
+date: 2011-05-17 00:00:00 -05:00
 default_executable: 
 dependencies: 
 - !ruby/object:Gem::Dependency 
@@ -36,18 +36,18 @@ dependencies:
   type: :development
   version_requirements: *id002
 - !ruby/object:Gem::Dependency 
-  name: activesupport
+  name: activerecord
   prerelease: false
   requirement: &id003 !ruby/object:Gem::Requirement 
     none: false
     requirements: 
     - - ">="
       - !ruby/object:Gem::Version 
-        version: 2.3.4
-  type: :runtime
+        version: "3"
+  type: :development
   version_requirements: *id003
 - !ruby/object:Gem::Dependency 
-  name: amatch
+  name: mysql
   prerelease: false
   requirement: &id004 !ruby/object:Gem::Requirement 
     none: false
@@ -55,19 +55,63 @@ dependencies:
     - - ">="
       - !ruby/object:Gem::Version 
         version: "0"
-  type: :runtime
+  type: :development
   version_requirements: *id004
 - !ruby/object:Gem::Dependency 
-  name: to_regexp
+  name: cohort_scope
   prerelease: false
   requirement: &id005 !ruby/object:Gem::Requirement 
+    none: false
+    requirements: 
+    - - ">="
+      - !ruby/object:Gem::Version 
+        version: "0"
+  type: :development
+  version_requirements: *id005
+- !ruby/object:Gem::Dependency 
+  name: weighted_average
+  prerelease: false
+  requirement: &id006 !ruby/object:Gem::Requirement 
+    none: false
+    requirements: 
+    - - ">="
+      - !ruby/object:Gem::Version 
+        version: "0"
+  type: :development
+  version_requirements: *id006
+- !ruby/object:Gem::Dependency 
+  name: activesupport
+  prerelease: false
+  requirement: &id007 !ruby/object:Gem::Requirement 
+    none: false
+    requirements: 
+    - - ">="
+      - !ruby/object:Gem::Version 
+        version: "3"
+  type: :runtime
+  version_requirements: *id007
+- !ruby/object:Gem::Dependency 
+  name: amatch
+  prerelease: false
+  requirement: &id008 !ruby/object:Gem::Requirement 
+    none: false
+    requirements: 
+    - - ">="
+      - !ruby/object:Gem::Version 
+        version: "0"
+  type: :runtime
+  version_requirements: *id008
+- !ruby/object:Gem::Dependency 
+  name: to_regexp
+  prerelease: false
+  requirement: &id009 !ruby/object:Gem::Requirement 
     none: false
     requirements: 
     - - ">="
       - !ruby/object:Gem::Version 
         version: 0.0.3
   type: :runtime
-  version_requirements: *id005
+  version_requirements: *id009
 description: Create dictionaries that link rows between two tables using loose matching (string similarity) by default and tight matching (regexp) by request.
 email: 
 - seamus@abshere.net
@@ -104,6 +148,7 @@ files:
 - examples/icao-bts.xls
 - lib/loose_tight_dictionary.rb
 - lib/loose_tight_dictionary/blocking.rb
+- lib/loose_tight_dictionary/cached_result.rb
 - lib/loose_tight_dictionary/identity.rb
 - lib/loose_tight_dictionary/result.rb
 - lib/loose_tight_dictionary/score.rb
@@ -114,6 +159,7 @@ files:
 - loose_tight_dictionary.gemspec
 - test/helper.rb
 - test/test_blocking.rb
+- test/test_cache.rb
 - test/test_identity.rb
 - test/test_loose_tight_dictionary.rb
 - test/test_loose_tight_dictionary_convoluted.rb.disabled
@@ -149,6 +195,7 @@ summary: Allows iterative development of dictionaries for big data sets.
 test_files: 
 - test/helper.rb
 - test/test_blocking.rb
+- test/test_cache.rb
 - test/test_identity.rb
 - test/test_loose_tight_dictionary.rb
 - test/test_loose_tight_dictionary_convoluted.rb.disabled