redis-textsearch 0.1.0

data/README.rdoc

@@ -0,0 +1,93 @@
+= Redis::TextSearch - Use Redis to perform text search from any type of class
+
+This gem implements an extremely fast text search using Redis, based on the patterns from
+James Gray's {lists and sets in Redis post}[http://blog.grayproductions.net/articles/lists_and_sets_in_redis]
+as well as Antirez's {text search gist}[http://gist.github.com/120067]. You can use it with any type of class,
+whether it be ActiveRecord, DataMapper, MongoRecord, or even a class having nothing to do with an ORM.
+
+This is not intended to be the most full-featured text search available.  For that, look into
+{Sphinx}[http://www.sphinxsearch.com/], {Solr}[http://lucene.apache.org/solr/], or
+{other alternatives}[http://en.wikipedia.org/wiki/Full_text_search#Open_Source_projects].
+This gem is designed to (a) be extremely fast and (b) handle any (or multiple) data stores.
+
+The only requirement this gem has is that your class must provide an +id+ instance method
+which returns the ID for that instance. ActiveRecord, DataMapper, and MongoRecord all have +id+ methods which
+are known to be suitable. Since "ID" can be any data type, you can even write an +id+ method of your own
+that just returns a string, or an MD5 of a filename, or something else unique. 
+
+== Installation
+
+    gem install gemcutter
+    gem tumble
+    gem install redis-textsearch
+
+== Initialization
+    
+If on Rails, config/initializers/redis.rb is a good place for this:
+
+    require 'redis'
+    require 'redis/text_search'
+    Redis::TextSearch.redis = Redis.new(:host => 'localhost', :port => 6379)
+
+== Example
+
+Model class:
+
+    class Post < ActiveRecord::Base
+      include Redis::TextSearch
+
+      text_index :title, :minlength => 2
+      text_index :tags, :exact => true
+
+      # Using AR callback (you can call update_text_indexes anywhere on your instance)
+      after_save do |r|
+        r.update_text_indexes
+      end
+      after_destroy do |r|
+        r.delete_text_indexes
+      end
+    end
+
+Create posts:
+
+    Post.create(:title => "All About Bacon", :tags => "chef nontechnical")
+    Post.create(:title => "All About Bacon - Part 2", :tags => "chef nontechnical")
+    Post.create(:title => "Homemade Belgian Waffles", :tags => "chef nontechnical")
+    Post.create(:title => "Using Redis with Ruby", :tags => "technical ruby redis")
+    Post.create(:title => "Installing Redis on Linux", :tags => "technical redis linux")
+    Post.create(:title => "Chef Deployment Recipes", :tags => "ruby howto chef")
+
+Then search for them:
+
+    @posts = Post.text_search('bacon')   # 2 results
+    @posts = Post.text_search('chef')    # 4 results (tags and title)
+    @posts = Post.text_search('technical', 'ruby')   # AND search (1 result)
+    @posts = Post.text_search('chef', :fields => :tags)   # 3 results (only search tags)
+
+You can pass options through to the +find+ method:
+
+    @posts = Post.text_search('redis', :order => 'updated_at desc')
+    @posts = Post.text_search('redis', :select => 'id,title', :limit => 50)
+
+And do pagination (adapted from +will_paginate+):
+
+    @posts = Post.text_search('redis', :page => 1, :per_page => 10)
+    @posts = Post.text_search('redis', :page => 1)  # uses class.per_page like will_paginate
+
+You can also specify specific fields to search as a hash: 
+
+    @posts = Post.text_search(:tags => 'chef')  # 4 results
+    @posts = Post.text_search(:tags => ['technical','ruby'])  # AND (1 result)
+    @posts = Post.text_search(:tags => 'redis', :title => 'linux')  # 1 result
+    @posts = Post.text_search(:title => 'chef') # 1 result
+
+Note that if you need to pass options to +find+ AND search specific fields, the first
+hash must be in brackets:
+
+    @posts = Post.text_search({:tags => 'chef', :title => 'deployment'},
+                              :order => 'updated_at desc')
+
+== Author
+
+Copyright (c) 2009 {Nate Wiger}[http://nate.wiger.org].  All Rights Reserved.
+Released under the {Artistic License}[http://www.opensource.org/licenses/artistic-license-2.0.php].

data/lib/redis/text_search/collection.rb

@@ -0,0 +1,146 @@
+module Redis::TextSearch
+  # = Invalid page number error
+  # This is an ArgumentError raised in case a page was requested that is either
+  # zero or negative number. You should decide how do deal with such errors in
+  # the controller.
+  #
+  # If you're using Rails 2, then this error will automatically get handled like
+  # 404 Not Found. The hook is in "will_paginate.rb":
+  #
+  #   ActionController::Base.rescue_responses['Redis::TextSearch::InvalidPage'] = :not_found
+  #
+  # If you don't like this, use your preffered method of rescuing exceptions in
+  # public from your controllers to handle this differently. The +rescue_from+
+  # method is a nice addition to Rails 2.
+  #
+  # This error is *not* raised when a page further than the last page is
+  # requested. Use <tt>Redis::TextSearch::Collection#out_of_bounds?</tt> method to
+  # check for those cases and manually deal with them as you see fit.
+  class InvalidPage < ArgumentError
+    def initialize(page, page_num)
+      super "#{page.inspect} given as value, which translates to '#{page_num}' as page number"
+    end
+  end
+  
+  # = The key to pagination
+  # Arrays returned from paginating finds are, in fact, instances of this little
+  # class. You may think of Redis::TextSearch::Collection as an ordinary array with
+  # some extra properties. Those properties are used by view helpers to generate
+  # correct page links.
+  #
+  # Redis::TextSearch::Collection also assists in rolling out your own pagination
+  # solutions: see +create+.
+  # 
+  # If you are writing a library that provides a collection which you would like
+  # to conform to this API, you don't have to copy these methods over; simply
+  # make your plugin/gem dependant on the "mislav-will_paginate" gem:
+  #
+  #   gem 'mislav-will_paginate'
+  #   require 'will_paginate/collection'
+  #   
+  #   # Redis::TextSearch::Collection is now available for use
+  class Collection < Array
+    attr_reader :current_page, :per_page, :total_entries, :total_pages
+
+    # Arguments to the constructor are the current page number, per-page limit
+    # and the total number of entries. The last argument is optional because it
+    # is best to do lazy counting; in other words, count *conditionally* after
+    # populating the collection using the +replace+ method.
+    def initialize(page, per_page, total = nil)
+      @current_page = page.to_i
+      raise InvalidPage.new(page, @current_page) if @current_page < 1
+      @per_page = per_page.to_i
+      raise ArgumentError, "`per_page` setting cannot be less than 1 (#{@per_page} given)" if @per_page < 1
+      
+      self.total_entries = total if total
+    end
+
+    # Just like +new+, but yields the object after instantiation and returns it
+    # afterwards. This is very useful for manual pagination:
+    #
+    #   @entries = Redis::TextSearch::Collection.create(1, 10) do |pager|
+    #     result = Post.find(:all, :limit => pager.per_page, :offset => pager.offset)
+    #     # inject the result array into the paginated collection:
+    #     pager.replace(result)
+    #
+    #     unless pager.total_entries
+    #       # the pager didn't manage to guess the total count, do it manually
+    #       pager.total_entries = Post.count
+    #     end
+    #   end
+    #
+    # The possibilities with this are endless. For another example, here is how
+    # Redis::TextSearch used to define pagination for Array instances:
+    #
+    #   Array.class_eval do
+    #     def paginate(page = 1, per_page = 15)
+    #       Redis::TextSearch::Collection.create(page, per_page, size) do |pager|
+    #         pager.replace self[pager.offset, pager.per_page].to_a
+    #       end
+    #     end
+    #   end
+    #
+    # The Array#paginate API has since then changed, but this still serves as a
+    # fine example of Redis::TextSearch::Collection usage.
+    def self.create(page, per_page, total = nil)
+      pager = new(page, per_page, total)
+      yield pager
+      pager
+    end
+
+    # Helper method that is true when someone tries to fetch a page with a
+    # larger number than the last page. Can be used in combination with flashes
+    # and redirecting.
+    def out_of_bounds?
+      current_page > total_pages
+    end
+
+    # Current offset of the paginated collection. If we're on the first page,
+    # it is always 0. If we're on the 2nd page and there are 30 entries per page,
+    # the offset is 30. This property is useful if you want to render ordinals
+    # side by side with records in the view: simply start with offset + 1.
+    def offset
+      (current_page - 1) * per_page
+    end
+
+    # current_page - 1 or nil if there is no previous page
+    def previous_page
+      current_page > 1 ? (current_page - 1) : nil
+    end
+
+    # current_page + 1 or nil if there is no next page
+    def next_page
+      current_page < total_pages ? (current_page + 1) : nil
+    end
+    
+    # sets the <tt>total_entries</tt> property and calculates <tt>total_pages</tt>
+    def total_entries=(number)
+      @total_entries = number.to_i
+      @total_pages   = (@total_entries / per_page.to_f).ceil
+    end
+
+    # This is a magic wrapper for the original Array#replace method. It serves
+    # for populating the paginated collection after initialization.
+    #
+    # Why magic? Because it tries to guess the total number of entries judging
+    # by the size of given array. If it is shorter than +per_page+ limit, then we
+    # know we're on the last page. This trick is very useful for avoiding
+    # unnecessary hits to the database to do the counting after we fetched the
+    # data for the current page.
+    #
+    # However, after using +replace+ you should always test the value of
+    # +total_entries+ and set it to a proper value if it's +nil+. See the example
+    # in +create+.
+    def replace(array)
+      result = super
+      
+      # The collection is shorter then page limit? Rejoice, because
+      # then we know that we are on the last page!
+      if total_entries.nil? and length < per_page and (current_page == 1 or length > 0)
+        self.total_entries = offset + length
+      end
+
+      result
+    end
+  end
+end

data/lib/redis/text_search.rb

@@ -0,0 +1,289 @@
+# Redis::TextSearch - Use Redis to add text search to your app.
+class Redis
+  #
+  # Redis::TextSearch enables high-performance text search in your app using
+  # Redis.  You can perform text search on any type of data store or ORM.
+  #
+  module TextSearch
+    class NoFinderMethod < StandardError; end
+    class BadTextIndex < StandardError; end
+
+    DEFAULT_EXCLUDE_LIST = %w(a an and as at but by for in into of on onto to the)
+
+    dir = File.expand_path(__FILE__.sub(/\.rb$/,''))
+    autoload :Collection, File.join(dir, 'collection')
+
+    class << self
+      def redis=(conn) @redis = conn end
+      def redis
+        @redis ||= $redis || raise(NotConnected, "Redis::TextSearch.redis not set to a Redis.new connection")
+      end
+
+      def included(klass)
+        klass.instance_variable_set('@redis', @redis)
+        klass.instance_variable_set('@text_indexes', {})
+        klass.instance_variable_set('@text_search_find', nil)
+        klass.instance_variable_set('@text_index_exclude_list', DEFAULT_EXCLUDE_LIST)
+        klass.send :include, InstanceMethods
+        klass.extend ClassMethods
+        klass.guess_text_search_find
+        class << klass
+          define_method(:per_page) { 30 } unless respond_to?(:per_page)
+        end
+      end
+    end
+
+    # These class methods are added to the class when you include Redis::TextSearch.
+    module ClassMethods
+      attr_accessor :redis
+      attr_reader :text_indexes
+      
+      # Words to exclude from text indexing.  By default, includes common
+      # English prepositions like "a", "an", "the", "and", "or", etc.
+      # This is an accessor to an array, so you can use += or << to add to it.
+      # See the constant +DEFAULT_EXCLUDE_LIST+ for the default list.
+      attr_accessor :text_index_exclude_list
+
+      # Set the Redis prefix to use. Defaults to model_name
+      def prefix=(prefix) @prefix = prefix end
+      def prefix #:nodoc:
+        @prefix ||= self.name.to_s.
+          sub(%r{(.*::)}, '').
+          gsub(/([A-Z]+)([A-Z][a-z])/,'\1_\2').
+          gsub(/([a-z\d])([A-Z])/,'\1_\2').
+          downcase
+      end
+      
+      def field_key(name, id) #:nodoc:
+        "#{prefix}:#{id}:#{name}"
+      end
+      
+      # This is called when the class is imported, and uses reflection to guess
+      # how to retrieve records.  You can override it by explicitly defining a
+      # +text_search_find+ class method that takes an array of IDs as an argument.
+      def guess_text_search_find
+        if defined?(ActiveRecord::Base) and is_a?(ActiveRecord::Base)
+          instance_eval <<-EndMethod
+            def text_search_find(ids, options)
+              all(options.merge(:conditions => {:#{primary_key} => ids}))
+            end
+          EndMethod
+        elsif defined?(MongoRecord::Base) and is_a?(MongoRecord::Base)
+          instance_eval <<-EndMethod
+            def text_search_find(ids, options)
+              all(options.merge(:conditions => {:#{primary_key} => ids}))
+            end
+          EndMethod
+        elsif respond_to?(:get)
+          # DataMapper::Resource is an include, so is_a? won't work
+          instance_eval <<-EndMethod
+            def text_search_find(ids, options)
+              get(ids, options)
+            end
+          EndMethod
+        end
+      end
+
+      # Define fields to be indexed for text search.  To update the index, you must call
+      # update_text_indexes after record save or at the appropriate point.
+      def text_index(*args)
+        options = args.last.is_a?(Hash) ? args.pop : {}
+        options[:minlength] ||= 1
+        options[:split] ||= /\s+/
+        raise ArgumentError, "Must specify fields to index to #{self.name}.text_index" unless args.length > 0
+        args.each do |name|
+          @text_indexes[name.to_sym] = options.merge(:key => field_key(name, 'text_index'))
+        end
+      end
+
+      # Perform text search and return results from database. Options:
+      #
+      #   'string', 'string2'
+      #   :fields
+      #   :page
+      #   :per_page
+      def text_search(*args)
+        options = args.length > 1 && args.last.is_a?(Hash) ? args.pop : {}
+        fields = Array(options.delete(:fields) || @text_indexes.keys)
+        finder = options.delete(:finder)
+        unless finder
+          unless defined?(:text_search_find)
+            raise NoFinderMethod, "Could not detect how to find records; you must def text_search_find()"
+          end
+          finder = :text_search_find
+        end
+
+        #
+        # Assemble set names for our intersection.
+        # Accept two ways of doing search: either {:field => ['value','value'], :field => 'value'},
+        # or 'value','value', :fields => [:field, :field].  The first is an AND, the latter an OR.
+        #
+        ids = []
+        if args.empty?
+          raise ArgumentError, "Must specify search string(s) to #{self.name}.text_search"
+        elsif args.first.is_a?(Hash)
+          sets = []
+          args.first.each do |f,v|
+            sets += text_search_sets_for(f,v)
+          end
+          # Execute single intersection (AND)
+          ids = redis.set_intersect(*sets)
+        else
+          fields.each do |f|
+            sets = text_search_sets_for(f,args)
+            # Execute intersection per loop (OR)
+            ids += redis.set_intersect(*sets)
+          end
+        end
+
+        # Calculate pagination if applicable. Presence of :page indicates we want pagination.
+        # Adapted from will_paginate/finder.rb
+        if options.has_key?(:page)
+          page = options.delete(:page) || 1
+          per_page = options.delete(:per_page) || self.per_page
+          total = ids.length
+
+          Redis::TextSearch::Collection.create(page, per_page, total) do |pager|
+            # Convert page/per_page to limit/offset
+            options.merge!(:offset => pager.offset, :limit => pager.per_page)
+            pager.replace(send(finder, ids, options){ |*a| yield(*a) if block_given? })
+          end
+        else
+          # Execute finder directly
+          send(finder, ids, options)
+        end
+      end
+
+      # Filter and return self. Chainable.
+      def text_filter(field)
+        raise UnimplementedError
+      end
+
+      # Delete all text indexes for the given id.
+      def delete_text_indexes(id, *fields)
+        fields = @text_indexes.keys if fields.empty?
+        fields.each do |field|
+          redis.pipelined do |pipe|
+            text_indexes_for(id, field).each do |key|
+              pipe.srem(key, id)
+            end
+            pipe.del field_key("#{field}_indexes", id)
+          end
+        end
+      end
+
+      def text_indexes_for(id, field) #:nodoc:
+        (redis.get(field_key("#{field}_indexes", id)) || '').split(';')
+      end
+
+      def text_search_sets_for(field, values)
+        key = @text_indexes[field][:key]
+        Array(values).collect do |val|
+          str = val.downcase.gsub(/[^\w\s]+/,'').gsub(/\s+/, '.')  # can't have " " in Redis cmd string
+          "#{key}:#{str}"
+        end
+      end
+    end
+
+    module InstanceMethods #:nodoc:
+      def redis() self.class.redis end
+      def field_key(name) #:nodoc:
+        self.class.field_key(name, id)
+      end
+
+      # Retrieve the options for the given field
+      def text_index_options_for(field)
+        self.class.text_indexes[field] ||
+          raise(BadTextIndex, "No such text index #{field} in #{self.class.name}")
+      end
+      
+      # Retrieve the reverse-mapping of text indexes for a given field.  Designed
+      # as a utility method but maybe you will find it useful.
+      def text_indexes_for(field)
+        self.class.text_indexes_for(id, field)
+      end
+
+      # Retrieve all text indexes
+      def text_indexes
+        fields = self.class.text_indexes.keys
+        fields.collect{|f| text_indexes_for(f)}.flatten
+      end
+
+      # Update all text indexes for the given object.  Should be used in an +after_save+ hook
+      # or other applicable area, for example r.update_text_indexes.  Can pass an array of
+      # field names to restrict updates just to those fields.
+      def update_text_indexes(*fields)
+        fields = self.class.text_indexes.keys if fields.empty?
+        fields.each do |field|
+          options = self.class.text_indexes[field]
+          value = self.send(field)
+          return false if value.length < options[:minlength]  # too short to index
+          indexes = []
+
+          # If values is array, like :tags => ["a", "b"], use as-is
+          # Otherwise, split words on /\s+/ so :title => "Hey there" => ["Hey", "there"]
+          values = value.is_a?(Array) ? value : options[:split] ? value.split(options[:split]) : value
+          values.each do |val|
+            val.gsub!(/[^\w\s]+/,'')
+            val.downcase!
+            next if value.length < options[:minlength]
+            next if self.class.text_index_exclude_list.include? value
+            if options[:exact]
+              str = val.gsub(/\s+/, '.')  # can't have " " in Redis cmd string
+              indexes << "#{options[:key]}:#{str}"
+            else
+              len = options[:minlength]
+              while len < val.length
+                str = val[0..len].gsub(/\s+/, '.')  # can't have " " in Redis cmd string
+                indexes << "#{options[:key]}:#{str}"
+                len += 1
+              end
+            end
+          end
+
+          # Determine what, if anything, needs to be done.  If the indexes are unchanged,
+          # don't make any trips to Redis.  Saves tons of useless network calls.
+          old_indexes = text_indexes_for(field)
+          new_indexes = indexes - old_indexes
+          del_indexes = old_indexes - indexes
+
+          # No change, so skip
+          # puts "[#{field}] old=#{old_indexes.inspect} / idx=#{indexes.inspect} / new=#{new_indexes.inspect} / del=#{del_indexes.inspect}"
+          next if new_indexes.empty? and del_indexes.empty?
+
+          # Add new indexes
+          exec_pipelined_index_cmd(:sadd, new_indexes)
+
+          # Delete indexes no longer used
+          exec_pipelined_index_cmd(:srem, del_indexes)
+          
+          # Replace our reverse map of indexes
+          redis.set field_key("#{field}_indexes"), indexes.join(';')
+        end # fields.each
+      end
+
+      # Delete all text indexes that the object is a member of.  Should be used in
+      # an +after_destroy+ hook to remove the dead object.
+      def delete_text_indexes(*fields)
+        fields = self.class.text_indexes.keys if fields.empty?
+        fields.each do |field|
+          del_indexes = text_indexes_for(field)
+          exec_pipelined_index_cmd(:srem, del_indexes)
+          redis.del field_key("#{field}_indexes")
+        end
+      end
+
+      private
+      
+      def exec_pipelined_index_cmd(cmd, indexes)
+        return if indexes.empty?
+        redis.pipelined do |pipe|
+          indexes.each do |key|
+            pipe.send(cmd, key, id)
+          end
+        end
+      end
+      
+    end
+  end
+end

data/spec/redis_text_search_core_spec.rb

@@ -0,0 +1,143 @@
+
+require File.expand_path(File.dirname(__FILE__) + '/spec_helper')
+
+class Post
+  include Redis::TextSearch
+
+  text_index :title
+  text_index :tags, :exact => true
+
+  def self.text_search_find(ids, options)
+    options.empty? ? ids : [ids, options]
+  end
+
+  def self.first(ids, options)
+    
+  end
+
+  def initialize(attrib)
+    @attrib = attrib
+    @id = attrib[:id] || 1
+  end
+  def id; @id; end
+  def method_missing(name, *args)
+    @attrib[name] || super
+  end
+end
+
+TITLES = [
+  'Some plain text',
+  'More plain textstring comments',
+  'Come get somebody personal comments',
+  '*Welcome to Nate\'s new BLOG!!',
+]
+
+TAGS = [
+  ['personal', 'nontechnical'],
+  ['mysql', 'technical'],
+  ['gaming','technical']
+]
+
+
+describe Redis::TextSearch do
+  before :all do
+    @post  = Post.new(:title => TITLES[0], :tags => TAGS[0], :id => 1)
+    @post2 = Post.new(:title => TITLES[1], :tags => TAGS[1], :id => 2)
+    @post3 = Post.new(:title => TITLES[2], :tags => TAGS[2], :id => 3)
+
+    @post.delete_text_indexes
+    @post2.delete_text_indexes
+    Post.delete_text_indexes(3)
+  end
+
+  it "should define text indexes in the class" do
+    Post.text_indexes[:title][:key].should   == 'post:text_index:title'
+    Post.text_indexes[:tags][:key].should == 'post:text_index:tags'
+  end
+
+  it "should update text indexes correctly" do
+    @post.update_text_indexes
+    @post2.update_text_indexes
+
+    Post.redis.set_members('post:text_index:title:so').should == ['1']
+    Post.redis.set_members('post:text_index:title:som').should == ['1']
+    Post.redis.set_members('post:text_index:title:some').should == ['1']
+    Post.redis.set_members('post:text_index:title:pl').sort.should == ['1','2']
+    Post.redis.set_members('post:text_index:title:pla').sort.should == ['1','2']
+    Post.redis.set_members('post:text_index:title:plai').sort.should == ['1','2']
+    Post.redis.set_members('post:text_index:title:plain').sort.should == ['1','2']
+    Post.redis.set_members('post:text_index:title:te').sort.should == ['1','2']
+    Post.redis.set_members('post:text_index:title:tex').sort.should == ['1','2']
+    Post.redis.set_members('post:text_index:title:text').sort.should == ['1','2']
+    Post.redis.set_members('post:text_index:title:texts').should == ['2']
+    Post.redis.set_members('post:text_index:title:textst').should == ['2']
+    Post.redis.set_members('post:text_index:title:textstr').should == ['2']
+    Post.redis.set_members('post:text_index:title:textstri').should == ['2']
+    Post.redis.set_members('post:text_index:title:textstrin').should == ['2']
+    Post.redis.set_members('post:text_index:title:textstring').should == ['2']
+    Post.redis.set_members('post:text_index:tags:pe').should == []
+    Post.redis.set_members('post:text_index:tags:per').should == []
+    Post.redis.set_members('post:text_index:tags:pers').should == []
+    Post.redis.set_members('post:text_index:tags:perso').should == []
+    Post.redis.set_members('post:text_index:tags:person').should == []
+    Post.redis.set_members('post:text_index:tags:persona').should == []
+    Post.redis.set_members('post:text_index:tags:personal').should == ['1']
+    Post.redis.set_members('post:text_index:tags:no').should == []
+    Post.redis.set_members('post:text_index:tags:non').should == []
+    Post.redis.set_members('post:text_index:tags:nont').should == []
+    Post.redis.set_members('post:text_index:tags:nonte').should == []
+    Post.redis.set_members('post:text_index:tags:nontec').should == []
+    Post.redis.set_members('post:text_index:tags:nontech').should == []
+    Post.redis.set_members('post:text_index:tags:nontechn').should == []
+    Post.redis.set_members('post:text_index:tags:nontechni').should == []
+    Post.redis.set_members('post:text_index:tags:nontechnic').should == []
+    Post.redis.set_members('post:text_index:tags:nontechnica').should == []
+    Post.redis.set_members('post:text_index:tags:nontechnical').should == ['1']
+  end
+
+  it "should search text indexes and return records" do
+    Post.text_search('some').should == ['1']
+    @post3.update_text_indexes
+    Post.text_search('some').sort.should == ['1','3']
+    Post.text_search('plain').sort.should == ['1','2']
+    Post.text_search('plain','text').sort.should == ['1','2']
+    Post.text_search('plain','textstr').sort.should == ['2']
+    Post.text_search('some','TExt').sort.should == ['1']
+    Post.text_search('techNIcal').sort.should == ['2','3']
+    Post.text_search('nontechnical').sort.should == ['1']
+    Post.text_search('personal').sort.should == ['1','3']
+    Post.text_search('personAL', :fields => :tags).sort.should == ['1']
+    Post.text_search('PERsonal', :fields => [:tags]).sort.should == ['1']
+    Post.text_search('nontechnical', :fields => [:title]).sort.should == []
+  end
+
+  it "should pass options thru to find" do
+    Post.text_search('some', :order => 'updated_at desc').should == [['3','1'], {:order=>"updated_at desc"}]
+    Post.text_search('some', :select => 'id,username').should == [['3','1'], {:select => 'id,username'}]
+  end
+
+  it "should handle pagination" do
+    Post.text_search('some', :page => 1).should == [['3','1'], {:offset=>0, :limit=>30}]
+    Post.text_search('some', :page => 2, :per_page => 5).should == [['3','1'], {:offset=>5, :limit=>5}]
+    Post.text_search('some', :page => 15, :per_page => 3).should == [['3','1'], {:offset=>42, :limit=>3}]
+  end
+
+  it "should support a hash to the text_search method" do
+    Post.text_search(:tags => 'technical').sort.should == ['2','3']
+    Post.text_search(:tags => 'nontechnical').sort.should == ['1']
+    Post.text_search(:tags => 'technical', :title => 'plain').should == ['2']
+    Post.text_search(:tags => ['technical','MYsql'], :title => 'Mo').should == ['2']
+    Post.text_search(:tags => ['technical','MYsql'], :title => 'some').should == []
+    Post.text_search(:tags => 'technical', :title => 'comments').sort.should == ['2','3']
+  end
+
+  # MUST BE LAST!!!!!!
+  it "should delete text indexes" do
+    @post.delete_text_indexes
+    @post2.delete_text_indexes
+    Post.delete_text_indexes(3)
+    @post.text_indexes.should == []
+    @post2.text_indexes.should == []
+    @post3.text_indexes.should == []
+  end
+end

data/spec/spec_helper.rb

@@ -0,0 +1,5 @@
+$LOAD_PATH.unshift File.expand_path(File.dirname(__FILE__) + '/../lib')
+require 'redis'
+require 'redis/text_search'
+
+Redis::TextSearch.redis = Redis.new(:host => ENV['REDIS_HOST'], :port => ENV['REDIS_PORT'])

metadata

@@ -0,0 +1,69 @@
+--- !ruby/object:Gem::Specification 
+name: redis-textsearch
+version: !ruby/object:Gem::Version 
+  version: 0.1.0
+platform: ruby
+authors: 
+- Nate Wiger
+autorequire: 
+bindir: bin
+cert_chain: []
+
+date: 2009-12-03 00:00:00 -08:00
+default_executable: 
+dependencies: 
+- !ruby/object:Gem::Dependency 
+  name: redis
+  type: :runtime
+  version_requirement: 
+  version_requirements: !ruby/object:Gem::Requirement 
+    requirements: 
+    - - ">="
+      - !ruby/object:Gem::Version 
+        version: "0.1"
+    version: 
+description: Crazy fast text search using Redis. Works with any ORM or data store.
+email: nate@wiger.org
+executables: []
+
+extensions: []
+
+extra_rdoc_files: 
+- README.rdoc
+files: 
+- lib/redis/text_search/collection.rb
+- lib/redis/text_search.rb
+- spec/redis_text_search_core_spec.rb
+- spec/spec_helper.rb
+- README.rdoc
+has_rdoc: true
+homepage: http://github.com/nateware/redis-textsearch
+licenses: []
+
+post_install_message: 
+rdoc_options: 
+- --title
+- Redis::TextSearch -- Crazy fast text search using Redis
+require_paths: 
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement 
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      version: "0"
+  version: 
+required_rubygems_version: !ruby/object:Gem::Requirement 
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      version: "0"
+  version: 
+requirements: 
+- redis, v0.1 or greater
+rubyforge_project: redis-textsearch
+rubygems_version: 1.3.5
+signing_key: 
+specification_version: 3
+summary: Fast text search and word indexes using Redis
+test_files: []
+