pose 0.0.3 → 0.2

data/lib/generators/pose_generator.rb

@@ -6,7 +6,7 @@ class PoseGenerator < Rails::Generators::Base
   source_root File.expand_path('../templates', __FILE__)
   
   def create_migration_file
-    migration_template 'migration.rb', 'db/migrate/pose.rb'
+    migration_template 'migration.rb', 'db/migrate/add_pose_tables.rb'
   end
   
   def self.next_migration_number(path)

data/lib/generators/templates/migration.rb

@@ -1,4 +1,5 @@
-class PoseMigrations < ActiveRecord::Migration
+class AddPoseTables < ActiveRecord::Migration
+
   def self.up
     create_table "pose_assignments" do |t|
       t.integer "pose_word_id",               :null => false
@@ -6,14 +7,14 @@ class PoseMigrations < ActiveRecord::Migration
       t.string  "posable_type", :limit => 40, :null => false
     end
 
-    add_index "pose_assignments", :word_id
-    add_index "pose_assignments", :wordable_id
+    add_index "pose_assignments", :pose_word_id
+    add_index "pose_assignments", :posable_id
 
     create_table "pose_words" do |t|
       t.string "text", :limit => 80, :null => false
     end
 
-    add_index "words", :text
+    add_index "pose_words", :text
   end
 
   def self.down

data/lib/pose/railtie.rb

@@ -0,0 +1,11 @@
+require "rails/railtie"
+
+module Pose
+
+  class Railtie < Rails::Railtie
+    rake_tasks do
+      load "tasks/pose_tasks.rake"
+    end
+  end
+
+end

data/lib/pose/version.rb

@@ -1,3 +1,3 @@
 module Pose
-  VERSION = "0.0.3"
+  VERSION = "0.2"
 end

data/lib/pose.rb

@@ -1,3 +1,8 @@
+# Note (KG): Need to include the rake DSL here to prevent deprecation warnings in the Rakefile.
+require 'rake'
+include Rake::DSL
+
+
 # Polymorphic search for ActiveRecord objects.
 module Pose
   extend ActiveSupport::Concern
@@ -7,7 +12,7 @@ module Pose
   CONFIGURATION = { :search_in_tests => false }
 
   included do
-    has_many :pose_assignments, :as => :posable
+    has_many :pose_assignments, :as => :posable, :dependent => :delete_all
     has_many :pose_words, :through => :pose_assignments
 
     after_save :update_pose_index
@@ -32,7 +37,7 @@ module Pose
 
     # Removes this objects from the search index.
     def delete_pose_index
-      self.words.clear if perform_search?
+      self.pose_words.clear if perform_search?
     end
 
     # Helper method.
@@ -100,7 +105,8 @@ module Pose
     def root_word raw_word
       result = []
       raw_word_copy = raw_word[0..-1]
-      raw_word_copy.gsub! /[()*<>'",;]/, ' '
+      raw_word_copy.gsub! '%20', ' '
+      raw_word_copy.gsub! /[()*<>'",;\?\-\=&%#]/, ' '
       raw_word_copy.gsub! /\s+/, ' '
       raw_word_copy.split(' ').each do |word|
         if Pose.is_url?(word)
@@ -120,11 +126,12 @@ module Pose
     # Returns all objects matching the given query.
     #
     # @param [String] query
-    # @param [Array<Class>] classes
+    # @param (Class|[Array<Class>]) classes
+    # @param [Number?] limit Optional limit.
     #
     # @return [Hash<Class, ActiveRecord::Relation>]
-    def search query, classes
-      
+    def search query, classes, limit = nil
+
       # Turn 'classes' into an array.
       classes = [classes].flatten
       classes_names = classes.map &:name
@@ -134,13 +141,14 @@ module Pose
       query.split(' ').each do |query_word|
         current_word_classes_and_ids = {}
         classes.each { |clazz| current_word_classes_and_ids[clazz.name] = [] }
-        PoseAssignment.joins(:pose_word) \
-                      .where(:pose_words => {:text.matches => "#{query_word}%"},
-                             :posable_type => classes_names) \
-                      .each do |pose_assignment|
+        query = PoseAssignment.joins(:pose_word) \
+                              .where(:pose_words => {:text.matches => "#{query_word}%"},
+                                     :posable_type => classes_names)
+        query = query.limit(limit) if limit
+        query.each do |pose_assignment|
           current_word_classes_and_ids[pose_assignment.posable_type] << pose_assignment.posable_id
         end
-        
+
         current_word_classes_and_ids.each do |class_name, ids|
           if result_classes_and_ids.has_key? class_name
             result_classes_and_ids[class_name] = result_classes_and_ids[class_name] & ids
@@ -162,6 +170,12 @@ module Pose
           end
         end
       end
-    end    
+    end
   end
 end
+
+
+require 'pose/posifier'
+require "pose/railtie" if defined? Rails
+require 'pose_assignment'
+require 'pose_word'

data/lib/pose_assignment.rb

@@ -2,4 +2,28 @@
 class PoseAssignment < ActiveRecord::Base
   belongs_to :pose_word
   belongs_to :posable, :polymorphic => true
-end
+
+  # Removes all PoseAssignments for the given class.
+  def self.cleanup_class_index clazz
+    PoseAssignment.delete_all(['posable_type=?', clazz.name])
+  end
+
+  # Removes all PoseAssignments that aren't used anymore.
+  def self.cleanup_orphaned_pose_assignments
+    PoseAssignment.find_each(:include => [:posable, :pose_word], :batch_size => 5000) do |assignment|
+
+      # Delete the assignment if the posable object no longer exists.
+      if assignment.posable.nil?
+        puts "deleting assignment '#{assignment.id}' because the posable object no longer exists."
+        assignment.delete
+        next
+      end
+
+      # Delete the assignment if the PoseWord for it no longer exists.
+      if assignment.pose_word.nil?
+        puts "deleting assignment '#{assignment.id}' because its word no longer exists."
+        assignment.delete
+      end
+    end
+  end
+end

data/lib/pose_word.rb

@@ -1,4 +1,13 @@
 # A single word in the search index.
 class PoseWord < ActiveRecord::Base
   has_many :pose_assignments
-end
+
+  def self.remove_unused_words
+    PoseWord.find_each(:include => [:pose_assignments], :batch_size => 5000) do |pose_word|
+      if pose_word.pose_assignments.size == 0
+        pose_word.delete
+        next
+      end
+    end
+  end
+end

data/lib/tasks/pose_tasks.rake

@@ -1,18 +1,21 @@
+include Rake::DSL
 namespace :pose do
 
   desc "Deletes and recreates the search index for all instances of the given class."
-  task :reindex_all, [:class_name] => :environment do
-    clazz = Kernel.const_get class_name
+  task :reindex_all, [:class_name] => [:environment] do |t, args|
+    clazz = Kernel.const_get args.class_name
+    progress_bar = ProgressBar.new "  reindexing", clazz.count
     clazz.find_each do |instance|
       instance.update_pose_words
+      progress_bar.inc
     end
+    progress_bar.finish
   end
   
-  desc "Removes the search index for all instances of the given classes"
-  task :remove_from_index, [:class_name] => :environment do
-    clazz = Kernel.const_get class_name
-    clazz.find_each do |instance|
-      instance.delete_pose_words
-    end
+  desc "Removes the search index for all instances of the given classes."
+  task :remove_from_index, [:class_name] => :environment do |t, args|
+    clazz = Kernel.const_get args.class_name
+    PoseAssignment.cleanup_class_index clazz
+    puts "Search index for class #{clazz.name} deleted.\n\n"
   end
-end
+end

data/spec/factories.rb

@@ -0,0 +1,12 @@
+Factory.define :pose_word do |word|
+  word.text { Faker::Lorem.words(1).first }
+end
+
+Factory.define :pose_assignment do |assignment|
+  assignment.association :pose_word
+  assignment.association :posable, :factory => :posable_one
+end
+
+Factory.define :posable_one do |posable|
+  posable.text { Faker::Lorem.words(3).join ' ' }
+end

data/spec/pose_assignment_spec.rb

@@ -0,0 +1,48 @@
+require 'spec_helper'
+
+describe PoseAssignment do
+
+  before :each do
+    PoseAssignment.delete_all
+  end
+  
+  describe "cleanup_class_index" do
+    
+    before :each do
+      Factory :pose_assignment, :posable_id => 1, :posable_type => 'PosableOne'
+      Factory :pose_assignment, :posable_id => 2, :posable_type => 'PosableTwo'
+      PoseAssignment.cleanup_class_index PosableOne
+    end
+    
+    it "deletes all PoseAssignments for the given class" do
+      PoseAssignment.where(:posable_type => 'PosableOne').should have(0).items
+    end
+    
+    it "doesn't delete PoseAssignments for other classes" do
+      PoseAssignment.where(:posable_type => 'PosableTwo').should have(1).items
+    end
+  end
+    
+  describe "cleanup_orphaned_pose_assignments" do
+    
+    it "deletes the assignment if the posable object doesn't exist" do
+      Factory :pose_assignment, :posable_id => 2, :posable_type => 'PosableOne'
+      PoseAssignment.count.should > 0
+      PoseAssignment.cleanup_orphaned_pose_assignments
+      PoseAssignment.should have(0).items
+    end
+    
+    it "deletes the assignment if the pose_word doesn't exist" do
+      assignment = Factory :pose_assignment, :pose_word => nil, :pose_word_id => 27
+      PoseAssignment.cleanup_orphaned_pose_assignments
+      PoseAssignment.find_by_id(assignment.id).should be_nil
+    end
+    
+    it "doesn't delete the assignment if it is still used" do
+      assignment = Factory :pose_assignment
+      PoseAssignment.cleanup_orphaned_pose_assignments
+      PoseAssignment.find_by_id(assignment.id).should_not be_nil
+    end
+  end
+  
+end

data/spec/pose_spec.rb

@@ -1,57 +1,11 @@
+# encoding: utf-8
+
 require "spec_helper"
 
 ActiveRecord::Base.establish_connection(:adapter => "sqlite3", :database => ":memory:")
 
-class PosableOne < ActiveRecord::Base
-  posify { [text] }
-end
-
-class PosableTwo < ActiveRecord::Base
-  posify { [text] }
-end
-
-def setup_db
-  ActiveRecord::Schema.define(:version => 1) do
-
-    create_table 'posable_ones' do |t|
-      t.string 'text'
-    end
-
-    create_table 'posable_twos' do |t|
-      t.string 'text'
-    end
-
-    create_table "pose_assignments" do |t|
-      t.integer "pose_word_id",               :null => false
-      t.integer "posable_id",                 :null => false
-      t.string  "posable_type", :limit => 20, :null => false
-    end
-
-    create_table "pose_words" do |t|
-      t.string "text", :limit => 80, :null => false
-    end
-  end
-end
-
-def teardown_db
-  ActiveRecord::Base.connection.tables.each do |table|
-    ActiveRecord::Base.connection.drop_table(table)
-  end
-end
-
-
 describe Pose do
   subject { PosableOne.new }
-
-  before :all do
-    setup_db
-    Pose::CONFIGURATION[:search_in_tests] = true
-  end
-
-  after :all do
-    teardown_db
-    Pose::CONFIGURATION[:search_in_tests] = false
-  end
   
   before :each do
     PosableOne.delete_all
@@ -181,6 +135,69 @@ describe Pose do
     end
   end
 
+
+  describe 'root_word' do
+
+    it 'converts words into singular' do
+      Pose.root_word('bars').should eql(['bar'])
+    end
+
+    it 'removes special characters' do
+      Pose.root_word('(bar').should eql(['bar'])
+      Pose.root_word('bar)').should eql(['bar'])
+      Pose.root_word('(bar)').should eql(['bar'])
+      Pose.root_word('>foo').should eql(['foo'])
+      Pose.root_word('<foo').should eql(['foo'])
+      Pose.root_word('"foo"').should eql(['foo'])
+      Pose.root_word('"foo').should eql(['foo'])
+      Pose.root_word("'foo'").should eql(['foo'])
+      Pose.root_word("'foo's").should eql(['foo'])
+      Pose.root_word("foo?").should eql(['foo'])
+      Pose.root_word("foo!").should eql(['foo'])
+      Pose.root_word("foo/bar").should eql(['foo', 'bar'])
+      Pose.root_word("foo-bar").should eql(['foo', 'bar'])
+      Pose.root_word("foo--bar").should eql(['foo', 'bar'])
+      Pose.root_word("foo.bar").should eql(['foo', 'bar'])
+    end
+
+    it 'removes umlauts' do
+      Pose.root_word('fünf').should eql(['funf'])
+    end
+
+    it 'splits up numbers' do
+      Pose.root_word('11.2.2011').should eql(['11', '2', '2011'])
+      Pose.root_word('11-2-2011').should eql(['11', '2', '2011'])
+      Pose.root_word('30:4-5').should eql(['30', '4', '5'])
+    end
+
+    it 'converts into lowercase' do
+      Pose.root_word('London').should eql(['london'])
+    end
+
+    it "stores single-letter words" do
+      Pose.root_word('a b').should eql(['a', 'b'])
+    end
+
+    it "does't encode external URLs" do
+      Pose.root_word('http://web.com').should eql(['http', 'web', 'com'])
+    end
+
+    it "doesn't store empty words" do
+      Pose.root_word('  one two  ').should eql(['one', 'two'])
+    end
+
+    it "removes duplicates" do
+      Pose.root_word('one_one').should eql(['one'])
+      Pose.root_word('one one').should eql(['one'])
+    end
+    
+    it "splits up complex URLs" do
+      Pose.root_word('books?id=p7uyWPcVGZsC&dq=closure%20definitive%20guide&pg=PP1#v=onepage&q&f=false').should eql([
+        "book", "id", "p7uywpcvgzsc", "dq", "closure", "definitive", "guide", "pg", "pp1", "v", "onepage", "q", "f", "false"])
+    end
+  end
+
+
   describe 'search' do
     
     it 'works' do

data/spec/pose_word_spec.rb

@@ -0,0 +1,23 @@
+require 'spec_helper'
+
+describe PoseWord do
+  
+  before :all do
+    PoseWord.delete_all
+  end
+  
+  describe 'remove_unused_words' do
+
+    it 'removes unused words' do
+      Factory :pose_word
+      PoseWord.remove_unused_words
+      PoseWord.count.should == 0
+    end
+    
+    it "doesn't remove used words" do
+      snippet = Factory :posable_one
+      PoseWord.remove_unused_words
+      PoseWord.count.should > 0
+    end
+  end
+end

data/spec/spec_helper.rb

@@ -11,6 +11,9 @@ require 'pose_assignment.rb'
 require "pose/posifier.rb"
 require 'active_support/core_ext/string'
 require 'meta_where'
+require 'faker'
+require 'factory_girl'
+FactoryGirl.find_definitions
 
 # Configure Rails Environment
 ENV["RAILS_ENV"] = "test"
@@ -21,7 +24,6 @@ Rails = Hashie::Mash.new({:env => 'test'})
 
 
 # Load support files
-#Dir["#{File.dirname(__FILE__)}/support/**/*.rb"].each { |f| require f }
 Dir["#{File.dirname(__FILE__)}/support/**/*.rb"].each { |f| require f }
 #Dir[Rails.root.join("spec/support/**/*.rb")].each {|f| require f}
 
@@ -40,6 +42,15 @@ RSpec.configure do |config|
   # examples within a transaction, remove the following line or assign false
   # instead of true.
 #  config.use_transactional_fixtures = true
+
+  config.before :suite do
+    setup_db
+    Pose::CONFIGURATION[:search_in_tests] = true
+  end
+  
+  config.after :suite do 
+    teardown_db
+  end
 end
 
 # Verifies that a taggable object has the given tags.
@@ -58,3 +69,41 @@ RSpec::Matchers.define :have_pose_words do |expected|
     "expected that subject would have pose words [#{expected.join ', '}], but it has [#{texts.join ', '}]"
   end
 end
+
+
+class PosableOne < ActiveRecord::Base
+  posify { text }
+end
+
+class PosableTwo < ActiveRecord::Base
+  posify { text }
+end
+
+def setup_db
+  ActiveRecord::Schema.define(:version => 1) do
+
+    create_table 'posable_ones' do |t|
+      t.string 'text'
+    end
+
+    create_table 'posable_twos' do |t|
+      t.string 'text'
+    end
+
+    create_table "pose_assignments" do |t|
+      t.integer "pose_word_id",               :null => false
+      t.integer "posable_id",                 :null => false
+      t.string  "posable_type", :limit => 20, :null => false
+    end
+
+    create_table "pose_words" do |t|
+      t.string "text", :limit => 80, :null => false
+    end
+  end
+end
+
+def teardown_db
+  ActiveRecord::Base.connection.tables.each do |table|
+    ActiveRecord::Base.connection.drop_table(table)
+  end
+end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: pose
 version: !ruby/object:Gem::Version
-  version: 0.0.3
+  version: '0.2'
   prerelease: 
 platform: ruby
 authors:
@@ -9,11 +9,11 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2011-12-21 00:00:00.000000000Z
+date: 2012-01-16 00:00:00.000000000Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rails
-  requirement: &70188120999860 !ruby/object:Gem::Requirement
+  requirement: &70285165402800 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -21,10 +21,10 @@ dependencies:
         version: 3.0.0
   type: :runtime
   prerelease: false
-  version_requirements: *70188120999860
+  version_requirements: *70285165402800
 - !ruby/object:Gem::Dependency
   name: meta_where
-  requirement: &70188120999440 !ruby/object:Gem::Requirement
+  requirement: &70285165402380 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -32,10 +32,10 @@ dependencies:
         version: '0'
   type: :runtime
   prerelease: false
-  version_requirements: *70188120999440
+  version_requirements: *70285165402380
 - !ruby/object:Gem::Dependency
   name: rake
-  requirement: &70188120998960 !ruby/object:Gem::Requirement
+  requirement: &70285165401920 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -43,10 +43,43 @@ dependencies:
         version: '0'
   type: :runtime
   prerelease: false
-  version_requirements: *70188120998960
+  version_requirements: *70285165401920
+- !ruby/object:Gem::Dependency
+  name: ruby-progressbar
+  requirement: &70285165401500 !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :runtime
+  prerelease: false
+  version_requirements: *70285165401500
+- !ruby/object:Gem::Dependency
+  name: factory_girl
+  requirement: &70285165401080 !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :development
+  prerelease: false
+  version_requirements: *70285165401080
+- !ruby/object:Gem::Dependency
+  name: faker
+  requirement: &70285165400660 !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :development
+  prerelease: false
+  version_requirements: *70285165400660
 - !ruby/object:Gem::Dependency
   name: hashie
-  requirement: &70188120998500 !ruby/object:Gem::Requirement
+  requirement: &70285165400240 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -54,10 +87,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *70188120998500
+  version_requirements: *70285165400240
 - !ruby/object:Gem::Dependency
   name: rspec
-  requirement: &70188120997960 !ruby/object:Gem::Requirement
+  requirement: &70285165399820 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -65,10 +98,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *70188120997960
+  version_requirements: *70285165399820
 - !ruby/object:Gem::Dependency
   name: sqlite3
-  requirement: &70188120953600 !ruby/object:Gem::Requirement
+  requirement: &70285165399400 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -76,7 +109,7 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *70188120953600
+  version_requirements: *70285165399400
 description: Pose ('Polymorphic Search') allows fulltext search for ActiveRecord objects.
 email:
 - kevin.goslar@gmail.com
@@ -106,6 +139,7 @@ files:
 - lib/generators/pose_generator.rb
 - lib/generators/templates/migration.rb
 - lib/pose/posifier.rb
+- lib/pose/railtie.rb
 - lib/pose/version.rb
 - lib/pose.rb
 - lib/pose_assignment.rb
@@ -114,7 +148,10 @@ files:
 - MIT-LICENSE
 - Rakefile
 - README.markdown
+- spec/factories.rb
+- spec/pose_assignment_spec.rb
 - spec/pose_spec.rb
+- spec/pose_word_spec.rb
 - spec/spec_helper.rb
 homepage: http://github.com/kevgo/pose
 licenses: []
@@ -141,6 +178,9 @@ signing_key:
 specification_version: 3
 summary: A polymorphic, storage-system independent search engine for Ruby on Rails.
 test_files:
+- spec/factories.rb
+- spec/pose_assignment_spec.rb
 - spec/pose_spec.rb
+- spec/pose_word_spec.rb
 - spec/spec_helper.rb
 has_rdoc: