redis-search 0.2 → 0.3

data/README.markdown

@@ -6,23 +6,25 @@ High performance real-time search (Support Chinese), index in Redis for Rails ap
 
 * Real-time search
 * High performance
+* Segment words search and prefix match search
+* Support ActiveRecord, Mongoid and others ORM
 
 ## Requirements
 
-* Redis 2.2
-* Libmmseg
+* Ruby 1.9.2
+* Redis 2.2+
 
 ## Install
 
 in Rails application Gemfile
 
-		gem 'redis','2.1.1'
-		gem "rmmseg-cpp-huacnlee", "0.2.8"
-		gem 'redis-search', '0.1'
+	gem 'redis','2.1.1'
+	gem "rmmseg-cpp-huacnlee", "0.2.8"
+	gem 'redis-search', '0.3'
 
 install bundlers
 
-		$ bundle install
+    $ bundle install
 
 ## Configure
 
@@ -30,9 +32,11 @@ create file in: config/initializers/redis_search.rb
 
     require "redis_search"
     redis = Redis.new(:host => "127.0.0.1",:port => "6379")
-    redis.select("app_name.search")
+    # change redis database to 3, you need use a special database for search feature.
+    redis.select(3)
     RedisSearch.configure do |config|
-     config.redis = redis
+      config.redis = redis
+      config.complete_max_length = 100
     end
 
 ## Usage
@@ -56,13 +60,35 @@ bind RedisSearch callback event, it will to rebuild search indexes when data cre
         self.category.name
       end
     end
+    
+    class User
+      include Mongoid::Document
+      include RedisSearch
+      
+      field :name
+      field :tagline
+      field :email
+      
+      redis_search_index(:title_field => :name,
+                         :prefix_index_enable => true,
+                         :ext_fields => [:email,:tagline])
+    end
 
-    # GET /searchs?q=title
     class SearchController < ApplicationController
+      # GET /searchs?q=title
       def index
         RedisSearch::Search.query(params[:q], :type => "Post")
       end
+      
+      # GET /search_users?q=j
+      def search_users
+        RedisSearch::Search.complete(params[:q], :type => "Post")
+      end
     end
+
+## Benchmark test
+
+* [https://gist.github.com/1150933](https://gist.github.com/1150933)
     
 ## Demo
 

data/lib/redis_search/base.rb

@@ -4,6 +4,7 @@ module RedisSearch
   module ClassMethods
     def redis_search_index(options = {})
       title_field = options[:title_field] || :title
+      prefix_index_enable = options[:prefix_index_enable] || false
       ext_fields = options[:ext_fields] || []
       class_eval %(
         def redis_search_ext_fields(ext_fields)
@@ -18,8 +19,11 @@ module RedisSearch
         def redis_search_index_create
           s = Search.new(:title => self.#{title_field}, :id => self.id, 
                           :exts => self.redis_search_ext_fields(#{ext_fields}), 
-                          :type => self.class.to_s)
+                          :type => self.class.to_s,
+                          :prefix_index_enable => #{prefix_index_enable})
           s.save
+          # release s
+          s = nil
         end
 
         before_destroy :redis_search_index_remove

data/lib/redis_search/config.rb

@@ -4,11 +4,20 @@ module RedisSearch
   end
   
   class Config
-    attr_accessor :redis, :debug
+    # Redis 
+    attr_accessor :redis
+    # Debug toggle
+    attr_accessor :debug
+    # config for max length of content with RedisSearch:Search.complete method，default 100
+    # Please change this with your real data length, short is fast
+    # For example: You use complete search for your User model name field, and the "name" as max length in 15 chars, then you can set here to 15
+    # warring! The long content will can't be found, if the config length less than real content.
+    attr_accessor :complete_max_length
     
     def initialize
       self.debug = false
       self.redis = nil
+      self.complete_max_length = 100
     end
   end
 end

data/lib/redis_search/search.rb

@@ -2,7 +2,7 @@
 require "rmmseg"
 module RedisSearch  
   class Search
-    attr_accessor :type, :title, :id, :exts
+    attr_accessor :type, :title, :id, :exts, :prefix_index_enable
     def initialize(options = {})
       self.exts = []
       options.keys.each do |k|
@@ -52,15 +52,19 @@ module RedisSearch
       return if words.blank?
       words.each do |word|
         next if not Search.word?(word)
-        save_zindex(word)
         key = Search.mk_sets_key(self.type,word)
         RedisSearch.config.redis.sadd(key, self.id)
       end
+      
+      # 建立前最索引
+      if prefix_index_enable
+        save_prefix_index(self.title)
+      end
     end
   
-    def save_zindex(word)
-      return if not Search.word?(word)
-      word = word.downcase
+    def save_prefix_index(title)
+      return if not Search.word?(title)
+      word = title.downcase
       key = Search.mk_complete_key(self.type)
       (1..(word.length)).each do |l|
         prefix = word[0...l]
@@ -96,15 +100,17 @@ module RedisSearch
       limit = options[:limit] || 10 
 
       prefix_matchs = []
-      rangelen = 100 # This is not random, try to get replies < MTU size
+      # This is not random, try to get replies < MTU size
+      rangelen = RedisSearch.config.complete_max_length
       prefix = w.downcase
       key = Search.mk_complete_key(type)
+      
       start = RedisSearch.config.redis.zrank(key,prefix)
-
       return [] if !start
       count = limit
+      max_range = start+(rangelen*limit)-1
+      range = RedisSearch.config.redis.zrange(key,start,max_range)
       while prefix_matchs.length <= count
-        range = RedisSearch.config.redis.zrange(key,start,start+rangelen-1)
         start += rangelen
         break if !range or range.length == 0
         range.each {|entry|
@@ -117,12 +123,25 @@ module RedisSearch
             prefix_matchs << entry[0...-1]
           end
         }
+        range = range[start..max_range]
       end
       words = []
       words = prefix_matchs.uniq.collect { |w| Search.mk_sets_key(type,w) }
-      ids = RedisSearch.config.redis.sunion(*words)
+      if words.length > 1
+        temp_store_key = "tmpsunionstore:#{words.join("+")}"   
+        if !RedisSearch.config.redis.exists(temp_store_key)
+          # 将多个词语组合对比，得到并集，并存入临时区域   
+          RedisSearch.config.redis.sunionstore(temp_store_key,*words)
+          # 将临时搜索设为1天后自动清除
+          RedisSearch.config.redis.expire(temp_store_key,86400)
+        end
+        # 根据需要的数量取出 ids
+        ids = RedisSearch.config.redis.sort(temp_store_key,:limit => [0,limit])
+      else
+        ids = RedisSearch.config.redis.sort(words.first,:limit => [0,limit])
+      end
       return [] if ids.blank?
-      hmget(type,ids, :limit => limit)
+      hmget(type,ids)
     end
 
     # Search items, this will split words by Libmmseg
@@ -142,14 +161,26 @@ module RedisSearch
       sort_field = options[:sort_field] || "id"
       words = words.collect { |w| Search.mk_sets_key(type,w) }
       return result if words.blank?
-      ids = RedisSearch.config.redis.sinter(*words)
-      hmget(type,ids, :limit => limit, :sort_field => sort_field)
+      temp_store_key = "tmpinterstore:#{words.join("+")}"
+      if words.length > 1
+        if !RedisSearch.config.redis.exists(temp_store_key)
+          # 将多个词语组合对比，得到交集，并存入临时区域
+          RedisSearch.config.redis.sinterstore(temp_store_key,*words)
+          # 将临时搜索设为1天后自动清除
+          RedisSearch.config.redis.expire(temp_store_key,86400)
+        end
+        # 根据需要的数量取出 ids
+        ids = RedisSearch.config.redis.sort(temp_store_key,:limit => [0,limit])
+      else
+        # 根据需要的数量取出 ids
+        ids = RedisSearch.config.redis.sort(words.first,:limit => [0,limit])
+      end
+      hmget(type,ids, :sort_field => sort_field)
     end
   
     private
       def self.hmget(type, ids, options = {})
         result = []
-        limit = options[:limit] || 10
         sort_field = options[:sort_field] || "id"
         return result if ids.blank?
         RedisSearch.config.redis.hmget(type,*ids).each do |r|
@@ -159,15 +190,7 @@ module RedisSearch
             Search.warn("Search.query failed: #{e}")
           end
         end
-        items = sort_result(result, type, sort_field)
-        items = items[0..limit-1] if items.length > limit
-        items
-      end
-  
-      def self.sort_result(items, type, sort_field)
-        return items if items.blank?
-        items = items.sort { |x,y| y[sort_field] <=> x[sort_field] }
-        items
+        result
       end
   end
 end

metadata

@@ -1,104 +1,74 @@
---- !ruby/object:Gem::Specification 
+--- !ruby/object:Gem::Specification
 name: redis-search
-version: !ruby/object:Gem::Version 
-  hash: 15
+version: !ruby/object:Gem::Version
+  version: '0.3'
   prerelease: 
-  segments: 
-  - 0
-  - 2
-  version: "0.2"
 platform: ruby
-authors: 
+authors:
 - Jason Lee
 autorequire: 
 bindir: bin
 cert_chain: []
-
-date: 2011-08-17 00:00:00 +08:00
-default_executable: 
-dependencies: 
-- !ruby/object:Gem::Dependency 
+date: 2011-08-17 00:00:00.000000000Z
+dependencies:
+- !ruby/object:Gem::Dependency
   name: rmmseg-cpp-huacnlee
-  prerelease: false
-  requirement: &id001 !ruby/object:Gem::Requirement 
+  requirement: &2159797520 !ruby/object:Gem::Requirement
     none: false
-    requirements: 
+    requirements:
     - - ~>
-      - !ruby/object:Gem::Version 
-        hash: 7
-        segments: 
-        - 0
-        - 2
-        - 8
+      - !ruby/object:Gem::Version
         version: 0.2.8
   type: :runtime
-  version_requirements: *id001
-- !ruby/object:Gem::Dependency 
-  name: redis
   prerelease: false
-  requirement: &id002 !ruby/object:Gem::Requirement 
+  version_requirements: *2159797520
+- !ruby/object:Gem::Dependency
+  name: redis
+  requirement: &2159797040 !ruby/object:Gem::Requirement
     none: false
-    requirements: 
+    requirements:
     - - ~>
-      - !ruby/object:Gem::Version 
-        hash: 9
-        segments: 
-        - 2
-        - 1
-        - 1
+      - !ruby/object:Gem::Version
         version: 2.1.1
   type: :runtime
-  version_requirements: *id002
-description: High performance real-time search (Support Chinese), index in Redis for Rails application.
-email: 
+  prerelease: false
+  version_requirements: *2159797040
+description: High performance real-time search (Support Chinese), index in Redis for
+  Rails application.
+email:
 - huacnlee@gmail.com
 executables: []
-
 extensions: []
-
 extra_rdoc_files: []
-
-files: 
+files:
 - lib/redis_search/base.rb
 - lib/redis_search/config.rb
 - lib/redis_search/search.rb
 - lib/redis_search.rb
 - README.markdown
-has_rdoc: true
 homepage: http://github.com/huacnlee/redis-search
 licenses: []
-
 post_install_message: 
 rdoc_options: []
-
-require_paths: 
+require_paths:
 - lib
-required_ruby_version: !ruby/object:Gem::Requirement 
+required_ruby_version: !ruby/object:Gem::Requirement
   none: false
-  requirements: 
-  - - ">="
-    - !ruby/object:Gem::Version 
-      hash: 3
-      segments: 
-      - 0
-      version: "0"
-required_rubygems_version: !ruby/object:Gem::Requirement 
+  requirements:
+  - - ! '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+required_rubygems_version: !ruby/object:Gem::Requirement
   none: false
-  requirements: 
-  - - ">="
-    - !ruby/object:Gem::Version 
-      hash: 23
-      segments: 
-      - 1
-      - 3
-      - 6
+  requirements:
+  - - ! '>='
+    - !ruby/object:Gem::Version
       version: 1.3.6
 requirements: []
-
 rubyforge_project: 
-rubygems_version: 1.6.2
+rubygems_version: 1.8.6
 signing_key: 
 specification_version: 3
-summary: High performance real-time search (Support Chinese), index in Redis for Rails application.
+summary: High performance real-time search (Support Chinese), index in Redis for Rails
+  application.
 test_files: []
-