acts_as_indexed 0.6.3 → 0.6.4

data/CHANGELOG

@@ -1,11 +1,16 @@
+===0.6.4 [16th August 2010]
+ - Added starts-with query type [nilbus - Edward Anderson]
+ - Various fixes and improvements.
+ - Real names given for all contributors.
+
 ===0.6.3 [5th July 2010]
- - index file path can now be definited as a Pathname as well as an array. [parndt]
- - Can now define which records are indexed and which are not via an :if proc. [madpilot]
- - Lots of tidying up. [parndt]
- - Rails 3 fixes. [myabc]
+ - index file path can now be definited as a Pathname as well as an array. [parndt - Philip Arndt]
+ - Can now define which records are indexed and which are not via an :if proc. [madpilot - Myles Eftos]
+ - Lots of tidying up. [parndt - Philip Arndt]
+ - Rails 3 fixes. [myabc - Alex Coles]
 
 ===0.6.2 [11th June 2010]
- - Now available as a Gem as well as the original plugin. [parndt - Thanks for doing most of the hard work.]
+ - Now available as a Gem as well as the original plugin. [parndt - Philip Arndt - Thanks for doing most of the hard work.]
 
 ===0.6.0 [10th June 2010]
  - Now supports Rails 3.x.x as well as Rails 2.x.x.
@@ -14,11 +19,11 @@
  - Deprecated find_with_index and will_paginate_search methods.
 
 ===0.5.3 [6th June 2010]
-- Now supports non-standard table names automatically. [nandalopes]
+- Now supports non-standard table names automatically. [nandalopes - Fernanda Lopes]
 
 ===0.5.2 [3rd May 2010]
-- Fix for Errno::ERANGE error related to certain Math.log calculations. [parndt]
-- Improved index detection in a shared-directory environment. [bob-p]
+- Fix for Errno::ERANGE error related to certain Math.log calculations. [parndt - Philip Arndt]
+- Improved index detection in a shared-directory environment. [bob-p - Thomas Pomfret]
 
 ===0.5.1 [11 June 2009]
 - Fixed Ruby 1.8.6 compatibility.
@@ -67,7 +72,7 @@
 
 ===0.3.0 [18 September 2007]
 - Minor bug fixes.
-- min_word_size now works properly, with quieries containing small words in 
+- min_word_size now works properly, with queries containing small words in 
   quotes or being preceded by a '+' symbol are now searched on.
 
 ===0.2.2 [06 September 2007]
@@ -93,4 +98,4 @@
 
 ===0.1 [31 August 2007]
 
-- Initial release.
+- Initial release.

data/README.rdoc

@@ -18,13 +18,13 @@ app with no dependencies and minimal setup.
 
 === Install
 
-== Rails 2.x.x
+==== Rails 2.x.x
   ./script/plugin install git://github.com/dougal/acts_as_indexed.git
 
-== Rails 3.x.x
+==== Rails 3.x.x
   rails plugin install git://github.com/dougal/acts_as_indexed.git
 
-=== As a Gem
+==== As a Gem
 Despite this being slightly against the the original ethos of the project,
 acts_as_indexed is now available as a Gem as several people have requested it.
 
@@ -32,6 +32,8 @@ acts_as_indexed is now available as a Gem as several people have requested it.
 
 Make sure to specify the Gem in your environment.rb file (Rails 2.x.x), or the Gemfile (Rails 3.x.x).
 
+==== No Git?
+
 If you don't have git installed, you can download the plugin from the GitHub
 page (http://github.com/dougal/acts_as_indexed) and unpack it into the
 <tt>vendor/plugins</tt> directory of your rails app.
@@ -88,7 +90,6 @@ an argument.
   # Chain it with any number of ActiveRecord methods and named_scopes.
   my_search_results = Post.public.with_query('my search query').find(:all, :limit => 10) # return the first 10 matches which are public.
 
-
 === Query Options
 
 The following query operators are supported:
@@ -97,6 +98,8 @@ The following query operators are supported:
 * NOT:: 'cat -dog' will find records matching 'cat' AND NOT 'dog'
 * INCLUDE:: 'cat +me' will find records matching 'cat' and 'me', even if 'me' is smaller than the +min_word_size+
 * "":: Quoted terms are matched as phrases. '"cat dog"' will find records matching the whole phrase. Quoted terms can be preceded by the NOT operator; 'cat -"big dog"' etc. Quoted terms can include words shorter than the +min_word_size+.
+* ^:: Terms that begin with ^ will match records that contain a word starting with the term. '^cat' will find matches containing 'cat', 'catapult', 'caterpillar' etc.
+* ^"":: A quoted term that begins with ^ matches any phrase that begin with this phrase. '^"cat d"' will find records matching the whole phrases "cat dog" and "cat dinner". This type of search is useful for autocomplete inputs.
 
 === Pagination
 
@@ -137,7 +140,7 @@ All of the above are most welcome. mailto:dougal.s@gmail.com
 
 == Credits
 
-Douglas F Shearer - http:douglasfshearer.com
+Douglas F Shearer - http://douglasfshearer.com
 
 
 == Future Releases
@@ -146,5 +149,5 @@ Future releases will be looking to add the following features:
 * Optional html scrubbing during indexing.
 * Ranking affected by field weightings.
 * Support for DataMapper, Sequel and the various MongoDB ORMs.
-* UTF-8 support. See the current solution here:
-https://gist.github.com/193903bb4e0d6e5debe1
+* UTF-8 support. See the current solution in the following Gist: 
+  https://gist.github.com/193903bb4e0d6e5debe1

data/VERSION

@@ -1 +1 @@
-0.6.3
+0.6.4

data/acts_as_indexed.gemspec

@@ -5,11 +5,11 @@
 
 Gem::Specification.new do |s|
   s.name = %q{acts_as_indexed}
-  s.version = "0.6.3"
+  s.version = "0.6.4"
 
   s.required_rubygems_version = Gem::Requirement.new(">= 0") if s.respond_to? :required_rubygems_version=
   s.authors = ["Douglas F Shearer"]
-  s.date = %q{2010-07-05}
+  s.date = %q{2010-08-16}
   s.description = %q{Acts As Indexed is a plugin which provides a pain-free way to add fulltext search to your Ruby on Rails app}
   s.email = %q{dougal.s@gmail.com}
   s.extra_rdoc_files = [

data/lib/acts_as_indexed.rb

@@ -95,9 +95,7 @@ module ActsAsIndexed #:nodoc:
       build_index unless aai_config.index_file.directory?
       index = SearchIndex.new(aai_config.index_file, aai_config.index_file_depth, aai_fields, aai_config.min_word_size, aai_config.if_proc)
       index.add_record(record)
-      index.save
       @query_cache = {}
-      true
     end
 
     # Removes the passed +record+ from the index. Clears the query cache.
@@ -105,11 +103,9 @@ module ActsAsIndexed #:nodoc:
     def index_remove(record)
       index = SearchIndex.new(aai_config.index_file, aai_config.index_file_depth, aai_fields, aai_config.min_word_size, aai_config.if_proc)
       # record won't be in index if it doesn't exist. Just return true.
-      return true unless index.exists?
+      return unless index.exists?
       index.remove_record(record)
-      index.save
       @query_cache = {}
-      true
     end
 
     # Updates the index.
@@ -119,12 +115,8 @@ module ActsAsIndexed #:nodoc:
     def index_update(record)
       build_index unless aai_config.index_file.directory?
       index = SearchIndex.new(aai_config.index_file, aai_config.index_file_depth, aai_fields, aai_config.min_word_size, aai_config.if_proc)
-      #index.remove_record(find(record.id))
-      #index.add_record(record)
       index.update_record(record,find(record.id))
-      index.save
       @query_cache = {}
-      true
     end
 
     # Finds instances matching the terms passed in +query+. Terms are ANDed by
@@ -153,12 +145,12 @@ module ActsAsIndexed #:nodoc:
       else
         logger.debug('Query held in cache.')
       end
-      return @query_cache[query].sort.reverse.map(&:first) if options[:ids_only] || @query_cache[query].empty?
+      return @query_cache[query].sort.reverse.map{|r| r.first} if options[:ids_only] || @query_cache[query].empty?
 
       # slice up the results by offset and limit
       offset = find_options[:offset] || 0
       limit = find_options.include?(:limit) ? find_options[:limit] : @query_cache[query].size
-      part_query = @query_cache[query].sort.reverse.slice(offset,limit).map(&:first)
+      part_query = @query_cache[query].sort.reverse.slice(offset,limit).map{|r| r.first}
 
       # Set these to nil as we are dealing with the pagination by setting
       # exactly what records we want.
@@ -179,7 +171,7 @@ module ActsAsIndexed #:nodoc:
            ranked_records[r] = @query_cache[query][r.id]
          end
 
-         ranked_records.to_a.sort_by{|a| a.last }.reverse.map(&:first)
+         ranked_records.to_a.sort_by{|a| a.last }.reverse.map{|r| r.first}
        end
       end
 
@@ -189,14 +181,9 @@ module ActsAsIndexed #:nodoc:
 
     # Builds an index from scratch for the current model class.
     def build_index
-      increment = 500
-      offset = 0
-      while (records = find(:all, :limit => increment, :offset => offset)).size > 0
-        #p "offset is #{offset}, increment is #{increment}"
-        index = SearchIndex.new(aai_config.index_file, aai_config.index_file_depth, aai_fields, aai_config.min_word_size, aai_config.if_proc)
-        offset += increment
+      index = SearchIndex.new(aai_config.index_file, aai_config.index_file_depth, aai_fields, aai_config.min_word_size, aai_config.if_proc)
+      find_in_batches({ :batch_size => 500 }) do |records|
         index.add_records(records)
-        index.save
       end
     end
 

data/lib/acts_as_indexed/search_atom.rb

@@ -14,8 +14,10 @@ module ActsAsIndexed #:nodoc:
     # W(T, D) = tf(T, D) * log ( DN / df(T))
     # weighting = frequency_in_this_record * log (total_number_of_records / number_of_matching_records)
 
-    def initialize
-      @records = {}
+    attr_reader :records
+
+    def initialize(records={})
+      @records = records
     end
 
     # Returns true if the given record is present.
@@ -49,6 +51,13 @@ module ActsAsIndexed #:nodoc:
       @records.delete(record_id)
     end
 
+    # Creates a new SearchAtom with the combined records from self and other
+    def +(other)
+      SearchAtom.new(@records.clone.merge!(other.records) { |key, _old, _new|
+                                                            _old + _new
+                                                          })
+    end
+
     # Returns at atom containing the records and positions of +self+ preceded by +former+
     # "former latter" or "big dog" where "big" is the former and "dog" is the latter.
     def preceded_by(former)
@@ -101,4 +110,4 @@ module ActsAsIndexed #:nodoc:
     end
 
   end
-end
+end

data/lib/acts_as_indexed/search_index.rb

@@ -22,19 +22,21 @@ module ActsAsIndexed #:nodoc:
     end
 
     # Adds +record+ to the index.
-    def add_record(record)
-      return @records_size unless @if_proc.call(record)
+    def add_record(record, no_save=false)
+      return unless @if_proc.call(record)
       condensed_record = condense_record(record)
       load_atoms(condensed_record)
       add_occurences(condensed_record,record.id)
       @records_size += 1
+      self.save unless no_save
     end
 
     # Adds multiple records to the index. Accepts an array of +records+.
     def add_records(records)
       records.each do |record|
-        add_record(record)
+        add_record(record, true)
       end
+      self.save
     end
 
     # Removes +record+ from the index.
@@ -44,29 +46,13 @@ module ActsAsIndexed #:nodoc:
       atoms.each do |a|
         @atoms[a].remove_record(record.id) if @atoms.has_key?(a)
         @records_size -= 1
-        #p "removing #{record.id} from #{a}"
       end
+      self.save
     end
 
     def update_record(record_new, record_old)
-      # Work out which atoms have modifications.
-      # Minimises loading and saving of partitions.
-      old_atoms = condense_record(record_old)
-      new_atoms = condense_record(record_new)
-
-      # Remove the old version from the appropriate atoms.
-      load_atoms(old_atoms)
-      old_atoms.each do |a|
-        @atoms[a].remove_record(record_new.id) if @atoms.has_key?(a)
-      end
-
-      if @if_proc.call(record_new)
-        # Add the new version to the appropriate atoms.
-        load_atoms(new_atoms)
-        # TODO: Make a version of this method that takes the
-        # atomised version of the record.
-        add_occurences(new_atoms, record_new.id)
-      end
+      remove_record(record_old)
+      add_record(record_new)
     end
 
     # Saves the current index partitions to the filesystem.
@@ -77,7 +63,6 @@ module ActsAsIndexed #:nodoc:
         (atoms_sorted[encoded_prefix(atom_name)] ||= {})[atom_name] = records
       end
       atoms_sorted.each do |e_p, atoms|
-        #p "Saving #{e_p}."
         @root.join(e_p.to_s).open("w+") do |f|
           Marshal.dump(atoms,f)
         end
@@ -95,29 +80,55 @@ module ActsAsIndexed #:nodoc:
     # Returns an array of IDs for records matching +query+.
     def search(query)
       return [] if query.nil?
-      load_atoms(cleanup_atoms(query))
+      load_options = { :start => true } if query[/\^/]
+      load_atoms(cleanup_atoms(query), load_options || {})
       queries = parse_query(query.dup)
       positive = run_queries(queries[:positive])
       positive_quoted = run_quoted_queries(queries[:positive_quoted])
       negative = run_queries(queries[:negative])
       negative_quoted = run_quoted_queries(queries[:negative_quoted])
+      starts_with = run_queries(queries[:starts_with], true)
+      start_quoted = run_quoted_queries(queries[:start_quoted], true)
 
-      if queries[:positive_quoted].any? && queries[:positive].any?
-        p = positive.delete_if{ |r_id,w| positive_quoted.exclude?(r_id) }
-        pq = positive_quoted.delete_if{ |r_id,w| positive.exclude?(r_id) }
-        results = p.merge(pq) { |r_id,old_val,new_val| old_val + new_val}
-      elsif queries[:positive].any?
-        results = positive
-      else
-        results = positive_quoted
+      results = {}
+
+      if queries[:start_quoted].any?
+        results = merge_query_results(results, start_quoted)
+      end
+      
+      if queries[:starts_with].any?
+        results = merge_query_results(results, starts_with)
+      end
+      
+      if queries[:positive_quoted].any?
+        results = merge_query_results(results, positive_quoted)
+      end
+      
+      if queries[:positive].any?
+        results = merge_query_results(results, positive)
       end
 
       negative_results = (negative.keys + negative_quoted.keys)
       results.delete_if { |r_id, w| negative_results.include?(r_id) }
-      #p results
       results
     end
-
+    
+    def merge_query_results(results1, results2)
+      # Return the other if one is empty.
+      return results1 if results2.empty?
+      return results2 if results1.empty?
+      
+      # Delete any records from results 1 that are not in results 2.
+      r1 = results1.delete_if{ |r_id,w| results2.exclude?(r_id) }
+      
+      
+      # Delete any records from results 2 that are not in results 1.
+      r2 = results2.delete_if{ |r_id,w| results1.exclude?(r_id) }
+      
+      # Merge the results by adding their respective scores.
+      r1.merge(r2) { |r_id,old_val,new_val| old_val + new_val}
+    end
+    
     # Returns true if the index root exists on the FS.
     #--
     # TODO: Make a private method called 'root_exists?' which checks for the root directory.
@@ -129,7 +140,6 @@ module ActsAsIndexed #:nodoc:
 
     # Gets the size file from the index.
     def load_record_size
-      #p "About to load #{@root.join('size')}"
       @root.join('size').open do |f|
         Marshal.load(f)
       end
@@ -144,7 +154,11 @@ module ActsAsIndexed #:nodoc:
 
     # Returns true if the given atom is present.
     def include_atom?(atom)
-      @atoms.has_key?(atom)
+      if atom.is_a? Regexp
+        @atoms.keys.grep(atom).any?
+      else
+        @atoms.has_key?(atom)
+      end
     end
 
     # Returns true if all the given atoms are present.
@@ -170,7 +184,6 @@ module ActsAsIndexed #:nodoc:
       condensed_record.each_with_index do |atom, i|
         add_atom(atom)
         @atoms[atom].add_position(record_id, i)
-        #p "adding #{record.id} to #{atom}"
       end
     end
 
@@ -197,6 +210,12 @@ module ActsAsIndexed #:nodoc:
 
     def parse_query(s)
 
+      # Find ^"foo bar".
+      start_quoted = []
+      while st_quoted = s.slice!(/\^\"[^\"]*\"/)
+        start_quoted << cleanup_atoms(st_quoted)
+      end
+
       # Find -"foo bar".
       negative_quoted = []
       while neg_quoted = s.slice!(/-\"[^\"]*\"/)
@@ -209,6 +228,12 @@ module ActsAsIndexed #:nodoc:
         positive_quoted << cleanup_atoms(pos_quoted)
       end
 
+      # Find ^foo.
+      starts_with = []
+      while st_with = s.slice!(/\^[\S]*/)
+        starts_with << cleanup_atoms(st_with).first
+      end
+
       # Find -foo.
       negative = []
       while neg = s.slice!(/-[\S]*/)
@@ -224,74 +249,106 @@ module ActsAsIndexed #:nodoc:
       # Find all other terms.
       positive += cleanup_atoms(s,true)
 
-      {:negative_quoted => negative_quoted, :positive_quoted => positive_quoted, :negative => negative, :positive => positive}
+      { :start_quoted => start_quoted,
+        :negative_quoted => negative_quoted,
+        :positive_quoted => positive_quoted,
+        :starts_with => starts_with,
+        :negative => negative,
+        :positive => positive }
     end
-
-    def run_queries(atoms)
+    
+    def run_queries(atoms, starts_with=false)
       results = {}
-      atoms.uniq.each do |atom|
+      atoms.each do |atom|
         interim_results = {}
-        if include_atom?(atom)
-          # Collect all the weightings for the current atom.
-          interim_results = @atoms[atom].weightings(@records_size)
-        end
-        if results.empty?
-          # If first time round, set results with initial weightings.
-          results = interim_results
-        else
-          # If second time round, add weightings together for records
-          # matching both atoms. Any matching only one are discarded.
-          rr = {}
-          interim_results.each do |r,w|
-            rr[r] = w + results[r] if results[r]
-          end
-          results = rr
-        end
+        
+        # If these atoms are to be run as 'starts with', make them a Regexp
+        # with a carat.
+        atom = /^#{atom}/ if starts_with
+
+        # Get the resulting matches, and break if none exist.
+        matches = get_atom_results(@atoms.keys, atom)
+        break if matches.nil?
+        
+        # Grab the record IDs and weightings.
+        interim_results = matches.weightings(@records_size)
+        
+        # Merge them with the results obtained already, if any.
+        results = results.empty? ? interim_results : merge_query_results(results, interim_results)
+        
+        break if results.empty?
+        
       end
-      #p results
       results
     end
-
-    def run_quoted_queries(quoted_atoms)
+    
+    def run_quoted_queries(quoted_atoms, starts_with=false)
       results = {}
       quoted_atoms.each do |quoted_atom|
         interim_results = {}
+        
+        break if quoted_atom.empty?
+        
+        # If these atoms are to be run as 'starts with', make the final atom a
+        # Regexp with a line-start anchor.
+        quoted_atom[-1] = /^#{quoted_atom.last}/ if starts_with
+        
+        # Little bit of memoization.
+        atoms_keys = @atoms.keys
+        
+        # Get the matches for the first atom.
+        matches = get_atom_results(atoms_keys, quoted_atom.first)
+        break if matches.nil?
+        
         # Check the index contains all the required atoms.
-        # match_atom = first_word_atom
         # for each of the others
         #   return atom containing records + positions where current atom is preceded by following atom.
         # end
-        # return records from final atom.
-        next unless include_atoms?(quoted_atom)
-        matches = @atoms[quoted_atom.first]
+        # Return records from final atom.
         quoted_atom[1..-1].each do |atom_name|
-          matches = @atoms[atom_name].preceded_by(matches)
+          interim_matches = get_atom_results(atoms_keys, atom_name)
+          if interim_matches.nil?
+            matches = nil
+            break
+          end
+          matches = interim_matches.preceded_by(matches)
         end
-        #results += matches.record_ids
 
+        break if matches.nil?
+        # Grab the record IDs and weightings.
         interim_results = matches.weightings(@records_size)
-        if results.empty?
-          results = interim_results
-        else
-          rr = {}
-          interim_results.each do |r,w|
-            rr[r] = w + results[r] if results[r]
-          end
-          #p results.class
-          results = rr
-        end
 
+        # Merge them with the results obtained already, if any.
+        results = results.empty? ? interim_results : merge_query_results(results, interim_results)
+        
+        break if results.empty?
+        
       end
       results
     end
 
-    def load_atoms(atoms)
+    def get_atom_results(atoms_keys, atom)
+      if atom.is_a? Regexp
+        matching_keys = atoms_keys.grep(atom)
+        results = SearchAtom.new
+        matching_keys.each do |key|
+          results += @atoms[key]
+        end
+        results
+      else
+        @atoms[atom]
+      end
+    end
+
+    def load_atoms(atoms, options={})
       # Remove duplicate atoms.
       # Remove atoms already in index.
       # Calculate prefixes.
       # Remove duplicate prefixes.
       atoms.uniq.reject{|a| include_atom?(a)}.collect{|a| encoded_prefix(a)}.uniq.each do |name|
-        if (atom_file = @root.join(name.to_s)).exist?
+        pattern = @root.join(name.to_s).to_s
+        pattern += '*' if options[:start]
+        Pathname.glob(pattern).each do |atom_file|
           atom_file.open do |f|
             @atoms.merge!(Marshal.load(f))
           end

data/test/abstract_unit.rb

@@ -43,7 +43,6 @@ class ActiveSupport::TestCase #:nodoc:
     # Makes a query to invoke the index build.
     assert_equal [], Post.find_with_index('badger')
     assert index_loc.exist?
-    true
   end
 
   def index_loc

data/test/acts_as_indexed_test.rb

@@ -1,4 +1,4 @@
-require File.expand_path("../abstract_unit", __FILE__)
+require File.dirname(__FILE__) + '/abstract_unit'
 
 class ActsAsIndexedTest < ActiveSupport::TestCase
   fixtures :posts
@@ -12,10 +12,10 @@ class ActsAsIndexedTest < ActiveSupport::TestCase
   def test_adds_to_index
     original_post_count = Post.count
     assert_equal [], Post.find_with_index('badger')
-    p = Post.new(:title => 'badger', :body => 'Thousands of them!')
-    assert p.save
+    post = Post.new(:title => 'badger', :body => 'Thousands of them!')
+    assert post.save
     assert_equal original_post_count+1, Post.count
-    assert_equal [p.id], Post.find_with_index('badger',{},{:ids_only => true})
+    assert_equal [post.id], Post.find_with_index('badger',{},{:ids_only => true})
   end
 
   def test_removes_from_index
@@ -66,10 +66,10 @@ class ActsAsIndexedTest < ActiveSupport::TestCase
   def test_scoped_simple_queries
     assert_equal [],  Post.find_with_index(nil)
     assert_equal [],  Post.with_query('')
-    assert_equal [5, 6],  Post.with_query('ship').map(&:id).sort
-    assert_equal [6],  Post.with_query('foo').map(&:id)
-    assert_equal [6],  Post.with_query('foo ship').map(&:id)
-    assert_equal [6],  Post.with_query('ship foo').map(&:id)
+    assert_equal [5, 6],  Post.with_query('ship').map{|r| r.id}.sort
+    assert_equal [6],  Post.with_query('foo').map{|r| r.id}
+    assert_equal [6],  Post.with_query('foo ship').map{|r| r.id}
+    assert_equal [6],  Post.with_query('ship foo').map{|r| r.id}
   end
 
   def test_negative_queries
@@ -80,9 +80,9 @@ class ActsAsIndexedTest < ActiveSupport::TestCase
   end
 
   def test_scoped_negative_queries
-    assert_equal [5, 6],  Post.with_query('crane').map(&:id).sort
-    assert_equal [5],  Post.with_query('crane -foo').map(&:id)
-    assert_equal [5],  Post.with_query('-foo crane').map(&:id)
+    assert_equal [5, 6],  Post.with_query('crane').map{|r| r.id}.sort
+    assert_equal [5],  Post.with_query('crane -foo').map{|r| r.id}
+    assert_equal [5],  Post.with_query('-foo crane').map{|r| r.id}
     assert_equal [],  Post.with_query('-foo') #Edgecase
   end
 
@@ -94,8 +94,8 @@ class ActsAsIndexedTest < ActiveSupport::TestCase
   end
 
   def test_scoped_quoted_queries
-    assert_equal [5],  Post.with_query('"crane ship"').map(&:id)
-    assert_equal [6],  Post.with_query('"crane big"').map(&:id)
+    assert_equal [5],  Post.with_query('"crane ship"').map{|r| r.id}
+    assert_equal [6],  Post.with_query('"crane big"').map{|r| r.id}
     assert_equal [],  Post.with_query('foo "crane ship"')
     assert_equal [],  Post.with_query('"crane badger"')
   end
@@ -106,10 +106,40 @@ class ActsAsIndexedTest < ActiveSupport::TestCase
   end
 
   def test_scoped_negative_quoted_queries
-    assert_equal [6],  Post.with_query('crane -"crane ship"').map(&:id)
+    assert_equal [6],  Post.with_query('crane -"crane ship"').map{|r| r.id}
     assert_equal [],  Post.with_query('-"crane big"') # Edgecase
   end
 
+  def test_start_queries
+    assert_equal [6,5],  Post.find_with_index('ship ^crane',{},{:ids_only => true})
+    assert_equal [6,5],  Post.find_with_index('^crane ship',{},{:ids_only => true})
+    assert_equal [6,5],  Post.find_with_index('^ship ^crane',{},{:ids_only => true})
+    assert_equal [6,5],  Post.find_with_index('^crane ^ship',{},{:ids_only => true})
+    assert_equal [6,5],  Post.find_with_index('^ship crane',{},{:ids_only => true})
+    assert_equal [6,5],  Post.find_with_index('crane ^ship',{},{:ids_only => true})
+    assert_equal [6,5],  Post.find_with_index('^crane',{},{:ids_only => true})
+    assert_equal [6,5],  Post.find_with_index('^cran',{},{:ids_only => true})
+    assert_equal [6,5],  Post.find_with_index('^cra',{},{:ids_only => true})
+    assert_equal [6,5,4],  Post.find_with_index('^cr',{},{:ids_only => true})
+    assert_equal [6,5,4,3,2,1], Post.find_with_index('^c',{},{:ids_only => true})
+    assert_equal [], Post.find_with_index('^notthere',{},{:ids_only => true})
+  end
+
+  def test_start_quoted_queries
+    assert_equal [6,5],  Post.find_with_index('^"crane" ship',{},{:ids_only => true})
+    assert_equal [6,5],  Post.find_with_index('ship ^"crane"',{},{:ids_only => true})
+    assert_equal [5],  Post.find_with_index('^"crane ship"',{},{:ids_only => true})
+    assert_equal [5],  Post.find_with_index('^"crane shi"',{},{:ids_only => true})
+    assert_equal [5],  Post.find_with_index('^"crane sh"',{},{:ids_only => true})
+    assert_equal [5],  Post.find_with_index('^"crane s"',{},{:ids_only => true})
+    assert_equal [6,5],  Post.find_with_index('^"crane "',{},{:ids_only => true})
+    assert_equal [6,5],  Post.find_with_index('^"crane"',{},{:ids_only => true})
+    assert_equal [6,5],  Post.find_with_index('^"cran"',{},{:ids_only => true})
+    assert_equal [6,5],  Post.find_with_index('^"cra"',{},{:ids_only => true})
+    assert_equal [6,5,4],  Post.find_with_index('^"cr"',{},{:ids_only => true})
+    assert_equal [6,5,4,3,2,1], Post.find_with_index('^"c"',{},{:ids_only => true})
+  end
+
   def test_find_options
     all_results = Post.find_with_index('crane',{},{:ids_only => true})
     first_result = Post.find_with_index('crane',{:limit => 1})

data/test/configuration_test.rb

@@ -1,4 +1,4 @@
-require File.expand_path("../abstract_unit", __FILE__)
+require File.dirname(__FILE__) + '/abstract_unit'
 include ActsAsIndexed
 
 class ConfigurationTest < ActiveSupport::TestCase

data/test/fixtures/posts.yml

@@ -1,37 +1,37 @@
-# Content generated using the random article feature on Wikipedia, http://en.wikipedia.org/wiki/Special:Random
-# Wikipedia content may be redistributed under the GNU Free Documentation License; http://en.wikipedia.org/wiki/Wikipedia:Text_of_the_GNU_Free_Documentation_License
-wikipedia_article_1:
-  id: 1
-  title: Body Count (video game)
-  body: Body Count is a 1994 First-person shooter for the Sega Mega Drive. It is one of the few games that make use of the Menacer light gun and the Mega Mouse. \n In the U.S. the game was released on the Sega Channel.
-  visible: 1
-
-wikipedia_article_2:
-  id: 2
-  title: Julien Ellis
-  body: Julien Ellis is a ice hockey goalie, born in Sorel, Quebec on January 27, 1986. He is currently 6'0" and weighs approximately 177 pounds. He wears number 34 and catches left. \n Julien played his entire junior hockey career in the QMJHL with the Shawinigan Cataractes. He was there from 2002 through the 2006 season, and played a total of 173 regular season games for them. During his time there, he recorded eight shutouts, as well as a career high .921 save percentage and 2.41 goals against average. \n Julien was chosen in round six of the 2004 NHL Entry Draft by the Vancouver Canucks, making him 189th overall pick and the 5th pick for Vancouver. \n His 2006-07 season was spent with the Victoria Salmon Kings of the ECHL, where he played 37 games and made 1,212 saves. Julien was called up the the Manitoba Moose of the AHL several times, where he played eight games.
-  visible: 1
-  
-wikipedia_article_3:
-  id: 3
-  title: Tuen Mun River
-  body: The Tuen Mun River is a river in Tuen Mun, New Territories, Hong Kong. It has many tributaries, with major ones coming from Lam Tei, Kau Keng Shan, Hung Shui Hang and Nai Wai. It flows south, splitting Tuen Mun into a west side and an east side. It eventually feeds into the Tuen Mun Typhoon Shelter, which is part of Castle Peak Bay.
-  visible: 0
-  
-wikipedia_article_4:
-  id: 4
-  title: So Happily Unsatisfied
-  body: So Happily Unsatisfied is an album that was recorded by the band Nine Days. It was intended to be the follow-up to their successful major-label debut, The Madding Crowd from 2000. The release date of the album was repeatedly delayed by Sony until the band was ultimately dropped. In the interim, the album had leaked onto the internet. The band has also put the whole album on their official website for the public to download.
-  visible: 1
-  
-wikipedia_article_5:
-  id: 5
-  title: SS Cornhusker State (T-ACS-6)
-  body: SS Cornhusker State (T-ACS-6) is a crane ship in ready reserve for the United States Navy. She is stationed at Cheatham Annex in Williamsburg, Virginia and is in ready reserve under the Military Sealift Command. The ship was named for the state of Nebraska, which is also known as the Cornhusker State. \n The ship was built by the Bath Iron Works. Her keel was laid on 27 November 1967, launched on 2 November 1968, and delivered 20 June 1969 as CV Stag Hound (MA 207). \n Stag Hound was acquired by the US Navy from the Maritime Administration in 1986 and was converted throughout 1987. She re-entered service as Cornhusker State on 12 March 1988, and has been in ready reserve since 1993.
-  visible: 0
-  
-article_similar_to_5:
-  id: 6
-  title: An article I made up by myself!
-  body: crane crane big ship foo
-  visible: 1
+# Content generated using the random article feature on Wikipedia, http://en.wikipedia.org/wiki/Special:Random
+# Wikipedia content may be redistributed under the GNU Free Documentation License; http://en.wikipedia.org/wiki/Wikipedia:Text_of_the_GNU_Free_Documentation_License
+wikipedia_article_1:
+  id: 1
+  title: Body Count (video game)
+  body: Body Count is a 1994 First-person shooter for the Sega Mega Drive. It is one of the few games that make use of the Menacer light gun and the Mega Mouse. \n In the U.S. the game was released on the Sega Channel.
+  visible: 1
+
+wikipedia_article_2:
+  id: 2
+  title: Julien Ellis
+  body: Julien Ellis is a ice hockey goalie, born in Sorel, Quebec on January 27, 1986. He is currently 6'0" and weighs approximately 177 pounds. He wears number 34 and catches left. \n Julien played his entire junior hockey career in the QMJHL with the Shawinigan Cataractes. He was there from 2002 through the 2006 season, and played a total of 173 regular season games for them. During his time there, he recorded eight shutouts, as well as a career high .921 save percentage and 2.41 goals against average. \n Julien was chosen in round six of the 2004 NHL Entry Draft by the Vancouver Canucks, making him 189th overall pick and the 5th pick for Vancouver. \n His 2006-07 season was spent with the Victoria Salmon Kings of the ECHL, where he played 37 games and made 1,212 saves. Julien was called up the the Manitoba Moose of the AHL several times, where he played eight games.
+  visible: 1
+  
+wikipedia_article_3:
+  id: 3
+  title: Tuen Mun River
+  body: The Tuen Mun River is a river in Tuen Mun, New Territories, Hong Kong. It has many tributaries, with major ones coming from Lam Tei, Kau Keng Shan, Hung Shui Hang and Nai Wai. It flows south, splitting Tuen Mun into a west side and an east side. It eventually feeds into the Tuen Mun Typhoon Shelter, which is part of Castle Peak Bay.
+  visible: 0
+  
+wikipedia_article_4:
+  id: 4
+  title: So Happily Unsatisfied
+  body: So Happily Unsatisfied is an album that was recorded by the band Nine Days. It was intended to be the follow-up to their successful major-label debut, The Madding Crowd from 2000. The release date of the album was repeatedly delayed by Sony until the band was ultimately dropped. In the interim, the album had leaked onto the internet. The band has also put the whole album on their official website for the public to download.
+  visible: 1
+  
+wikipedia_article_5:
+  id: 5
+  title: SS Cornhusker State (T-ACS-6)
+  body: SS Cornhusker State (T-ACS-6) is a crane ship in ready reserve for the United States Navy. She is stationed at Cheatham Annex in Williamsburg, Virginia and is in ready reserve under the Military Sealift Command. The ship was named for the state of Nebraska, which is also known as the Cornhusker State. \n The ship was built by the Bath Iron Works. Her keel was laid on 27 November 1967, launched on 2 November 1968, and delivered 20 June 1969 as CV Stag Hound (MA 207). \n Stag Hound was acquired by the US Navy from the Maritime Administration in 1986 and was converted throughout 1987. She re-entered service as Cornhusker State on 12 March 1988, and has been in ready reserve since 1993.
+  visible: 0
+  
+article_similar_to_5:
+  id: 6
+  title: An article I made up by myself!
+  body: crane crane big ship foo
+  visible: 1

data/test/schema.rb

@@ -1,7 +1,13 @@
-ActiveRecord::Schema.define :version => 0 do
-  create_table :posts, :force => true do |t|
-    t.column :title, :string
-    t.column :body, :text
-    t.column :visible, :boolean
-  end
-end
+ActiveRecord::Schema.define :version => 0 do
+  create_table :posts, :force => true do |t|
+    t.column :title, :string
+    t.column :body, :text
+    t.column :visible, :boolean
+  end
+
+  create_table :sources, :force => true do |t|
+    t.column :name, :string
+    t.column :url, :string
+    t.column :description, :text
+  end
+end

data/test/search_atom_test.rb

@@ -1,4 +1,4 @@
-require File.expand_path("../abstract_unit", __FILE__)
+require File.dirname(__FILE__) + '/abstract_unit'
 include ActsAsIndexed
 
 class SearchAtomTest < ActiveSupport::TestCase  
@@ -81,6 +81,18 @@ class SearchAtomTest < ActiveSupport::TestCase
     assert_in_delta(3.219, weightings[1], 2 ** -10)
     assert_in_delta(1.609, weightings[2], 2 ** -10)
   end
+
+  def test_adding_with_recursive_merge
+    sa0 = SearchAtom.new()
+    sa1 = SearchAtom.new({1=>[1]})
+    sa2 = SearchAtom.new({1=>[2], 2=>[3]})
+    
+    assert_equal (sa0 + sa1).records, {1=>[1]}
+    assert_equal (sa0 + sa2).records, {1=>[2], 2=>[3]}
+    
+    assert_equal (sa1 + sa2).records, {1=>[1,2], 2=>[3]}
+    assert_equal (sa2 + sa1).records, {1=>[2,1], 2=>[3]}
+  end
   
   private
   

data/test/search_index_test.rb

@@ -1,4 +1,4 @@
-require File.expand_path("../abstract_unit", __FILE__)
+require File.dirname(__FILE__) + '/abstract_unit'
 include ActsAsIndexed
 
 class SearchIndexTest < ActiveSupport::TestCase
@@ -35,8 +35,8 @@ class SearchIndexTest < ActiveSupport::TestCase
     search_index = build_search_index
     mock_records = ['record0', 'record1']
     
-    search_index.expects(:add_record).with('record0')
-    search_index.expects(:add_record).with('record1')
+    search_index.expects(:add_record).with('record0', true)
+    search_index.expects(:add_record).with('record1', true)
     
     search_index.add_records(mock_records)
   end

metadata

@@ -1,13 +1,13 @@
 --- !ruby/object:Gem::Specification 
 name: acts_as_indexed
 version: !ruby/object:Gem::Version 
-  hash: 1
+  hash: 15
   prerelease: false
   segments: 
   - 0
   - 6
-  - 3
-  version: 0.6.3
+  - 4
+  version: 0.6.4
 platform: ruby
 authors: 
 - Douglas F Shearer
@@ -15,7 +15,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2010-07-05 00:00:00 +01:00
+date: 2010-08-16 00:00:00 +01:00
 default_executable: 
 dependencies: []