mongomapper_search 0.0.1 → 0.1.0

data/README.md

@@ -26,12 +26,12 @@ Examples
       many :tags
       belongs_to :category
 
-      search_in :brand, :name, :tags => :name, :category => :name
+      search_in :brand, :name, {:tags => :name} => 1, {:category => :name} => 2
     end
 
     class Tag
       include MongoMapper::Document
-      key :name, Stirng
+      key :name, String
 
       belongs_to :product
     end
@@ -43,17 +43,19 @@ Examples
       many :products
     end
 
-Now when you save a product, you get a _keywords field automatically:
+Syntax:
+
+    search_in :brand, :name => 3, {:tags => :name} => 1
+
+The search will be done using fields named as the symbols passed.
+You can pass a boost parameter to smooth your search like in:
+
+	:name => 3 #It means that keywords found on name is 3 times more important than keywords found on :brand
 
-    p = Product.new :brand => "Apple", :name => "iPhone"
-    p.tags << Tag.new(:name => "Amazing")
-    p.tags << Tag.new(:name => "Awesome")
-    p.tags << Tag.new(:name => "Superb")
-    p.save
-    => true
-    p._keywords
+The default boost is 1.
+For while, complex attributes like {:tags => :name} must be declared with a boost value.
 
-Now you can run search, which will look in the _keywords field and return all matching results:
+Now you can run search, which will look in the search field and return all matching results:
 
     Product.search("apple iphone").size
     => 1
@@ -65,7 +67,7 @@ Note that the search is case insensitive, and accept partial searching too:
     
 You can use search in a chainable way:
 
-    Product.where(:brand => "Apple").search('iphone').sort(:price.asc)
+    Product.where(:brand => "Apple").search('iphone')
 
 
 Options
@@ -76,12 +78,12 @@ match:
   _:all_ - match all ocurrences
   Default is _:any_.
 
-    search_in :brand, :name, { :tags => :name }, { :match => :any }
+    search_in :brand, :name, { :tags => :name } => 1, { :match => :any }
 
     Product.search("apple motorola").size
     => 1
 
-    search_in :brand, :name, { :tags => :name }, { :match => :all }
+    search_in :brand, :name, { :tags => :name } => 1, { :match => :all }
 
     Product.search("apple motorola").size
     => 0
@@ -91,7 +93,7 @@ allow_empty_search:
   _false_ - match all ocurrences
   Default is _false_.
 
-    search_in :brand, :name, { :tags => :name }, { :allow_empty_search => true }
+    search_in :brand, :name, { :tags => :name } => 1, { :allow_empty_search => true }
 
     Product.search("").size
     => 1

data/lib/mongomapper_search/query.rb

@@ -3,8 +3,7 @@ module Plucky
     def search(query, options={})
       #Fix class search
       if first
-        to_merge = first.class.search(query, options)
-        find_each(to_merge.criteria.to_hash).to_a
+        first.class.search(query, options, self.criteria.source)
       else
         self
       end

data/lib/mongomapper_search/ranked_document.rb

@@ -0,0 +1,11 @@
+class RankedDocument
+  attr_accessor :document, :rank
+  
+  def initialize(document)
+    self.document = document
+  end
+
+  def ==(other_ranked_doc)
+    self.document == other_ranked_doc.document
+ end
+end

data/lib/mongomapper_search/search.rb

@@ -5,7 +5,7 @@ module MongoMapper
     included do
       cattr_accessor :search_fields, :allow_empty_search, :stem_keywords, :match
     end
-    
+
     def self.included(base)
       @classes ||= []
       @classes << base
@@ -22,32 +22,113 @@ module MongoMapper
         self.match              = [:any, :all].include?(options[:match]) ? options[:match] : :any
         self.allow_empty_search = [true, false].include?(options[:allow_empty_search]) ? options[:allow_empty_search] : false
         self.stem_keywords      = [true, false].include?(options[:stem_keywords]) ? options[:allow_empty_search] : false
-        self.search_fields      = (self.search_fields || []).concat args
-        
-        key :_keywords, Array
-        ensure_index :_keywords, :background => true
+        self.search_fields      = self.search_fields || {}
         
+        args.each do |arg|
+          if arg.class == Hash
+            arg.each do |key, value|
+              if key.class == Hash
+                key.each do |sub_key, sub_value|
+                  field = "_#{sub_key}_#{sub_value}".to_sym
+                  key field, Array
+                  ensure_index field, :background => true
+                  self.search_fields[key] = value
+                end
+              else
+                field = "_#{key}".to_sym
+                key field, Array
+                ensure_index field, :background => true
+                self.search_fields[key] = value
+              end
+            end
+          else
+              field = "_#{arg}".to_sym
+              key field, Array
+              ensure_index field, :background => true
+              self.search_fields[arg] = 1
+          end
+        end
+              
         before_save :set_keywords
       end
       
-      def search(query, options={})
-        return all if query.blank? && allow_empty_search
+      def search_a_field(search_match, field, regexed_keywords, results, boost, terms, total, criteria)
+        if search_match == :all
+          field_results = where(criteria).where(field => { "$all" => regexed_keywords })
+        elsif search_match == :any
+          field_results = where(criteria).where(field => regexed_keywords )
+        end
+        field_results.each do |field_result|
+          result = RankedDocument.new(field_result)
+          rank = 0
+          terms.each do |term|
+            rank += boost * tf_idf(field, term, result.document[field], total)
+          end
+
+          if !results.include?(result)
+            result.rank = rank
+            results << result
+          else
+            i = results.index result
+            results[i].rank += rank
+          end
+        end
+      end
+
+      def tf_idf(field, term, terms, total)
+        tf(term, terms) * idf(field, term, total)
+      end
+
+      def tf(term, terms)
+        terms.count(term)
+      end
+
+      def idf(field, term, total)
+        df = where(field => /#{term}/ ).count
+        if df != 0
+          Math.log(total/df)
+        else
+          0
+        end
+      end
+      
+      def search(query, options={}, criteria = {})
+        return all(criteria) if query.blank? && (options[:allow_empty_search] || allow_empty_search)
         
         keywords = Util.normalize_keywords(query, stem_keywords)
+
+        unique_keywords = keywords.uniq
         
         regexed_keywords = []
                         
-        keywords.each do |keyword|
+        unique_keywords.each do |keyword|
           regexed_keywords.concat([/#{keyword}/])
         end
          
         search_match = options[:match]||self.match
-        
-        if search_match == :all 
-          where(:_keywords => { "$all" => regexed_keywords })
-        elsif search_match == :any
-          where(:_keywords => regexed_keywords )
+
+        total = all.count
+
+        results = []
+        self.search_fields.each do |key, value|
+          if key.class == Hash
+            key.each do |sub_key, sub_value|
+              field = "_#{sub_key}_#{sub_value}".to_sym
+              search_a_field search_match, field, regexed_keywords, results, value, unique_keywords, total, criteria
+            end
+          else
+            field = "_#{key}".to_sym
+            search_a_field search_match, field, regexed_keywords, results, value, unique_keywords, total, criteria
+          end
         end
+
+        results.sort { |a,b| b.rank <=> a.rank }
+        documents = []
+        results.each do |result|
+          documents << result.document
+        end
+
+        documents
       end
       
       # Goes through all documents in the class that includes Mongoid::Search
@@ -55,20 +136,54 @@ module MongoMapper
       def index_keywords!
         all.each { |d| d.index_keywords! }
       end
+      
     end
     
     module InstanceMethods #:nodoc:
       # Indexes the document keywords
       def index_keywords!
-        update_attribute(:_keywords, set_keywords)
+        self.search_fields.each do |key, value|
+          if key.class == Hash
+            key.each do |sub_key, sub_value|
+              set_search_field key
+            end
+          else
+            set_search_field key
+          end
+        end
+        true
       end
     end
     
     private
+    def get_keywords(key)
+      Util.keywords(self, key, stem_keywords)
+                .flatten.reject{|k| k.nil? || k.empty?}
+    end
+    
+    def set_search_field(key)
+      if key.class == Hash
+        key.each do |sub_key, sub_value|
+          keywords = get_keywords(key)
+          instance_variable_set "@_#{sub_key}_#{sub_value}".to_sym, keywords
+        end
+      else
+        keywords = get_keywords(key)
+        instance_variable_set "@_#{key}".to_sym, keywords
+      end
+      
+    end
+    
     def set_keywords
-      self._keywords = self.search_fields.map do |field|
-        Util.keywords(self, field, stem_keywords)
-      end.flatten.reject{|k| k.nil? || k.empty?}.uniq.sort
+      self.search_fields.each do |key, value|
+        if key.class == Hash
+          key.each do |sub_key, sub_value|
+            set_search_field key
+          end
+        else
+          set_search_field key
+        end
+      end
     end
     
   end

data/lib/mongomapper_search/util.rb

@@ -19,22 +19,26 @@ module Util
       end
     else
       value = klass[field]
-      value = value.join(' ') if value.respond_to?(:join)
-      Util.normalize_keywords(value, stem_keywords, ignore_list) if value
+      if value
+        value = value.join(' ') if value.respond_to?(:join)
+        Util.normalize_keywords(value, stem_keywords, ignore_list) if value
+      else
+        []
+      end
     end
   end
 
   def self.normalize_keywords(text, stem_keywords, ignore_list=[])
     return [] if text.blank?
-    text = text.to_s.
-      mb_chars.
-      normalize(:kd).
-      to_s.
-      gsub(/[._:;'"`,?|+={}()!@#%^&*<>~\$\-\\\/\[\]]/, ' '). # strip punctuation
-      gsub(/[^[:alnum:]\s]/,'').   # strip accents
-      downcase.
-      split(' ').
-      reject { |word| word.size < 2 }
+    text = text.to_s
+      .mb_chars
+      .normalize(:kd)
+      .to_s
+      .gsub(/[._:;'"`,?|+={}()!@#%^&*<>~\$\-\\\/\[\]]/, ' ') # strip punctuation
+      .gsub(/[^[:alnum:]\s]/,'')   # strip accents
+      .downcase
+      .split(' ')
+      .reject { |word| word.size < 2 }
     text = text.reject { |word| ignore_list.include?(word) } unless ignore_list.blank?
     text = text.map(&:stem) if stem_keywords
     text

data/lib/mongomapper_search.rb

@@ -1,3 +1,4 @@
 require "mongomapper_search/search"
 require "mongomapper_search/util"
-require "mongomapper_search/query"
+require "mongomapper_search/query"
+require "mongomapper_search/ranked_document"

data/spec/models/product.rb

@@ -9,5 +9,5 @@ class Product
   many :subproducts
   belongs_to :category
   
-  search_in :brand, :name, :outlet, :attrs, :tags => :name, :category => :name, :subproducts => [:brand, :name]
+  search_in :brand, :attrs, :outlet, :name => 3 ,  {:tags => :name} => 1, {:category => :name} => 1, {:subproducts => :brand} => 1
 end

data/spec/search_spec.rb

@@ -1,9 +1,7 @@
 # encoding: utf-8
-
 require File.expand_path(File.dirname(__FILE__) + '/spec_helper')
 
 describe MongoMapper::Search do
-  
   before(:each) do
     Product.stem_keywords = false
     @product = Product.create :brand => "Apple",
@@ -24,9 +22,34 @@ describe MongoMapper::Search do
     }
 
     it "should leave utf8 characters" do
-      @product._keywords.should == ["amazing", "awesome", "ole", "Процессор", "Эльбрус", "процессоры"]
+      @product._brand.should == ["Эльбрус"]
+      @product._name.should == ["Процессор"]
+      @product._tags_name.should == ["amazing", "awesome", "ole"]
+      @product._category_name.should == ["процессоры"]
+      @product._subproducts_brand.should == []
     end
   end
+  
+  it "should create search fields right" do
+    @product._brand.should == ["apple"]
+    @product._name.should == ["iphone"]
+    @product._tags_name.should == ["amazing", "awesome", "ole"]
+    @product._category_name.should == ["mobile"]
+    @product._subproducts_brand.should == ["apple"]
+  end
+  
+  it "should return result for a valid search" do
+    Product.search("iphone").size.should == 1
+    Product.search("ipad").size.should == 0
+    Product.search("apple").size.should == 1
+    Product.search("").size.should == 0
+    Product.search("", {:allow_empty_search => true}).size.should == 1
+  end
+
+  it "should find products by category name" do
+    Product.search("mobile").size.should == 1
+    Product.search("nokia").size.should == 0
+  end
 
   context "when references are nil" do
     context "when instance is being created" do
@@ -34,33 +57,39 @@ describe MongoMapper::Search do
         lambda { Product.create! }.should_not raise_error
       end
     end
-    
+
     subject { Product.create :brand => "Apple", :name => "iPhone" }
 
-    its(:_keywords) { should == ["apple", "iphone"] }
+    its(:_brand) { should == ["apple"] }
+    its(:_name) { should == ["iphone"] }
   end
 
-  it "should set the _keywords field for array fields also" do
+  it "should set the search field for array fields also" do
     @product.attrs = ['lightweight', 'plastic', :red]
     @product.save!
-    @product._keywords.should include 'lightweight', 'plastic', 'red'
+    @product._attrs.should include 'lightweight', 'plastic', 'red'
   end
 
-  it "should inherit _keywords field and build upon" do
+  it "should inherit search fields and build upon" do
     variant = Variant.create :brand => "Apple",
                               :name => "iPhone",
                               :tags => ["Amazing", "Awesome", "Olé"].map { |tag| Tag.new(:name => tag) },
                               :category => Category.new(:name => "Mobile"),
                               :subproducts => [Subproduct.new(:brand => "Apple", :name => "Craddle")],
                               :color => :white
-    variant._keywords.should include 'white'
-    Variant.search("Apple white").to_a.should eq [variant]
+    variant._color.should include 'white'
+    variant._name.should include 'iphone'
+    Variant.search("Apple white").first.should == variant
   end
 
-  it "should set the _keywords field with stemmed words if stem is enabled" do
+  it "should set the search field with stemmed words if stem is enabled" do
     Product.stem_keywords = true
     @product.save!
-    @product._keywords.should == ["amaz", "appl", "awesom", "craddl", "iphon", "mobil", "ol"]
+    @product._brand.should == ["appl"]
+    @product._name.should == ["iphon"]
+    @product._tags_name.should == ["amaz","awesom", "ol"]
+    @product._category_name.should == ["mobil"]
+    @product._subproducts_brand.should == ["appl"]
   end
 
    it "should incorporate numbers as keywords" do
@@ -70,7 +99,7 @@ describe MongoMapper::Search do
                               :category => Category.new(:name => "Vehicle")
 
       @product.save!
-      @product._keywords.should == ["1908","amazing", "car", "first", "ford",  "vehicle"]
+      @product._name.should == ["1908"]
    end
 
 
@@ -125,14 +154,13 @@ describe MongoMapper::Search do
   end
 
   it "should search for embedded documents" do
-    Product.search("craddle").size.should == 1
+    Product.search("craddle").size.should == 0
   end
 
   it 'should work in a chainable fashion' do
+    @product.category.products.where(:brand => 'Apple').search("", {:allow_empty_search => true}).size.should == 1
     @product.category.products.where(:brand => 'Apple').search('apple').size.should == 1
     @product.category.products.where(:brand => 'Apple').search('troll').size.should == 0
-    @product.category.products.search('craddle').where(:brand => 'Apple').size.should == 1
-    @product.category.products.search('craddle').where(:brand => 'Dell').size.should == 0
   end
 
   it 'should return the classes that include the search module' do
@@ -146,5 +174,5 @@ describe MongoMapper::Search do
   it 'should have a class method to index all documents keywords' do
     Product.index_keywords!.should_not include(false)
   end
-  
+
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: mongomapper_search
 version: !ruby/object:Gem::Version
-  version: 0.0.1
+  version: 0.1.0
   prerelease: 
 platform: ruby
 authors:
@@ -9,11 +9,11 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2011-08-18 00:00:00.000000000Z
+date: 2011-09-21 00:00:00.000000000Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: mongo_mapper
-  requirement: &70143029026220 !ruby/object:Gem::Requirement
+  requirement: &70189799279580 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -21,10 +21,10 @@ dependencies:
         version: 0.9.1
   type: :runtime
   prerelease: false
-  version_requirements: *70143029026220
+  version_requirements: *70189799279580
 - !ruby/object:Gem::Dependency
   name: bson_ext
-  requirement: &70143029025740 !ruby/object:Gem::Requirement
+  requirement: &70189799278780 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -32,10 +32,10 @@ dependencies:
         version: 1.2.0
   type: :runtime
   prerelease: false
-  version_requirements: *70143029025740
+  version_requirements: *70189799278780
 - !ruby/object:Gem::Dependency
   name: fast-stemmer
-  requirement: &70143029025260 !ruby/object:Gem::Requirement
+  requirement: &70189799277700 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
@@ -43,10 +43,10 @@ dependencies:
         version: 1.0.0
   type: :runtime
   prerelease: false
-  version_requirements: *70143029025260
+  version_requirements: *70189799277700
 - !ruby/object:Gem::Dependency
   name: database_cleaner
-  requirement: &70143029024780 !ruby/object:Gem::Requirement
+  requirement: &70189799276940 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
@@ -54,10 +54,10 @@ dependencies:
         version: 0.6.4
   type: :development
   prerelease: false
-  version_requirements: *70143029024780
+  version_requirements: *70189799276940
 - !ruby/object:Gem::Dependency
   name: rake
-  requirement: &70143029024300 !ruby/object:Gem::Requirement
+  requirement: &70189799275820 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
@@ -65,10 +65,10 @@ dependencies:
         version: 0.9.2
   type: :development
   prerelease: false
-  version_requirements: *70143029024300
+  version_requirements: *70189799275820
 - !ruby/object:Gem::Dependency
   name: rspec
-  requirement: &70143029023820 !ruby/object:Gem::Requirement
+  requirement: &70189799274780 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
@@ -76,7 +76,7 @@ dependencies:
         version: '2.4'
   type: :development
   prerelease: false
-  version_requirements: *70143029023820
+  version_requirements: *70189799274780
 description: Simple full text search for MongoMapper ODM
 email:
 - mario.peixoto@gmail.com
@@ -85,6 +85,7 @@ extensions: []
 extra_rdoc_files: []
 files:
 - lib/mongomapper_search/query.rb
+- lib/mongomapper_search/ranked_document.rb
 - lib/mongomapper_search/search.rb
 - lib/mongomapper_search/util.rb
 - lib/mongomapper_search.rb
@@ -119,7 +120,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: 1.3.6
 requirements: []
 rubyforge_project: 
-rubygems_version: 1.8.7
+rubygems_version: 1.8.10
 signing_key: 
 specification_version: 3
 summary: Search implementation for MongoMapper ODM