hawkerb-thinking-sphinx 1.3.14

data/lib/thinking_sphinx/configuration.rb

@@ -0,0 +1,282 @@
+require 'erb'
+require 'singleton'
+
+module ThinkingSphinx
+  # This class both keeps track of the configuration settings for Sphinx and
+  # also generates the resulting file for Sphinx to use.
+  # 
+  # Here are the default settings, relative to RAILS_ROOT where relevant:
+  #
+  # config file::           config/#{environment}.sphinx.conf
+  # searchd log file::      log/searchd.log
+  # query log file::        log/searchd.query.log
+  # pid file::              log/searchd.#{environment}.pid
+  # searchd files::         db/sphinx/#{environment}/
+  # address::               127.0.0.1
+  # port::                  9312
+  # allow star::            false
+  # min prefix length::     1
+  # min infix length::      1
+  # mem limit::             64M
+  # max matches::           1000
+  # morphology::            nil
+  # charset type::          utf-8
+  # charset table::         nil
+  # ignore chars::          nil
+  # html strip::            false
+  # html remove elements::  ''
+  # searchd_binary_name::   searchd
+  # indexer_binary_name::   indexer
+  #
+  # If you want to change these settings, create a YAML file at
+  # config/sphinx.yml with settings for each environment, in a similar
+  # fashion to database.yml - using the following keys: config_file,
+  # searchd_log_file, query_log_file, pid_file, searchd_file_path, port,
+  # allow_star, enable_star, min_prefix_len, min_infix_len, mem_limit,
+  # max_matches, morphology, charset_type, charset_table, ignore_chars,
+  # html_strip, html_remove_elements, delayed_job_priority,
+  # searchd_binary_name, indexer_binary_name.
+  #
+  # I think you've got the idea.
+  # 
+  # Each setting in the YAML file is optional - so only put in the ones you
+  # want to change.
+  #
+  # Keep in mind, if for some particular reason you're using a version of
+  # Sphinx older than 0.9.8 r871 (that's prior to the proper 0.9.8 release),
+  # don't set allow_star to true.
+  # 
+  class Configuration
+    include Singleton
+    
+    SourceOptions = %w( mysql_connect_flags mysql_ssl_cert mysql_ssl_key
+      mysql_ssl_ca sql_range_step sql_query_pre sql_query_post 
+      sql_query_killlist sql_ranged_throttle sql_query_post_index unpack_zlib
+      unpack_mysqlcompress unpack_mysqlcompress_maxsize )
+    
+    IndexOptions  = %w( charset_table charset_type charset_dictpath docinfo
+      enable_star exceptions html_index_attrs html_remove_elements html_strip
+      index_exact_words ignore_chars inplace_docinfo_gap inplace_enable
+      inplace_hit_gap inplace_reloc_factor inplace_write_factor min_infix_len
+      min_prefix_len min_stemming_len min_word_len mlock morphology ngram_chars
+      ngram_len ondisk_dict overshort_step phrase_boundary phrase_boundary_step
+      preopen stopwords stopwords_step wordforms )
+    
+    CustomOptions = %w( disable_range )
+        
+    attr_accessor :searchd_file_path, :allow_star, :database_yml_file,
+      :app_root, :model_directories, :delayed_job_priority
+    
+    attr_accessor :source_options, :index_options
+    
+    attr_reader :environment, :configuration, :controller
+    
+    # Load in the configuration settings - this will look for config/sphinx.yml
+    # and parse it according to the current environment.
+    # 
+    def initialize(app_root = Dir.pwd)
+      self.reset
+    end
+    
+    def self.configure(&block)
+      yield instance
+      instance.reset(instance.app_root)
+    end
+    
+    def reset(custom_app_root=nil)
+      if custom_app_root
+        self.app_root = custom_app_root
+      else
+        self.app_root          = RAILS_ROOT if defined?(RAILS_ROOT)
+        self.app_root          = Merb.root  if defined?(Merb)
+        self.app_root        ||= app_root
+      end
+      
+      @configuration = Riddle::Configuration.new
+      @configuration.searchd.pid_file   = "#{self.app_root}/log/searchd.#{environment}.pid"
+      @configuration.searchd.log        = "#{self.app_root}/log/searchd.log"
+      @configuration.searchd.query_log  = "#{self.app_root}/log/searchd.query.log"
+      
+      @controller = Riddle::Controller.new @configuration,
+        "#{self.app_root}/config/#{environment}.sphinx.conf"
+      
+      self.address              = "127.0.0.1"
+      self.port                 = 9312
+      self.database_yml_file    = "#{self.app_root}/config/database.yml"
+      self.searchd_file_path    = "#{self.app_root}/db/sphinx/#{environment}"
+      self.allow_star           = false
+      self.model_directories    = ["#{app_root}/app/models/"] +
+        Dir.glob("#{app_root}/vendor/plugins/*/app/models/")
+      self.delayed_job_priority = 0
+      
+      self.source_options  = {}
+      self.index_options   = {
+        :charset_type => "utf-8"
+      }
+      
+      parse_config
+      
+      self
+    end
+    
+    def self.environment
+      Thread.current[:thinking_sphinx_environment] ||= (
+        defined?(Merb) ? Merb.environment : ENV['RAILS_ENV']
+      ) || "development"
+    end
+    
+    def environment
+      self.class.environment
+    end
+    
+    # Generate the config file for Sphinx by using all the settings defined and
+    # looping through all the models with indexes to build the relevant
+    # indexer and searchd configuration, and sources and indexes details.
+    #
+    def build(file_path=nil)
+      file_path ||= "#{self.config_file}"
+      
+      @configuration.indexes.clear
+      
+      ThinkingSphinx.context.indexed_models.each do |model|
+        model = model.constantize
+        model.define_indexes
+        @configuration.indexes.concat model.to_riddle
+      end
+      
+      open(file_path, "w") do |file|
+        file.write @configuration.render
+      end
+    end
+    
+    def address
+      @address
+    end
+    
+    def address=(address)
+      @address = address
+      @configuration.searchd.address = address
+    end
+    
+    def port
+      @port
+    end
+    
+    def port=(port)
+      @port = port
+      @configuration.searchd.port = port
+    end
+    
+    def pid_file
+      @configuration.searchd.pid_file
+    end
+    
+    def pid_file=(pid_file)
+      @configuration.searchd.pid_file = pid_file
+    end
+    
+    def searchd_log_file
+      @configuration.searchd.log
+    end
+    
+    def searchd_log_file=(file)
+      @configuration.searchd.log = file
+    end
+    
+    def query_log_file
+      @configuration.searchd.query_log
+    end
+    
+    def query_log_file=(file)
+      @configuration.searchd.query_log = file
+    end
+    
+    def config_file
+      @controller.path
+    end
+    
+    def config_file=(file)
+      @controller.path = file
+    end
+    
+    def bin_path
+      @controller.bin_path
+    end
+    
+    def bin_path=(path)
+      @controller.bin_path = path
+    end
+    
+    def searchd_binary_name
+      @controller.searchd_binary_name
+    end
+    
+    def searchd_binary_name=(name)
+      @controller.searchd_binary_name = name
+    end
+    
+    def indexer_binary_name
+      @controller.indexer_binary_name
+    end
+    
+    def indexer_binary_name=(name)
+      @controller.indexer_binary_name = name
+    end
+    
+    def client
+      client = Riddle::Client.new address, port
+      client.max_matches = configuration.searchd.max_matches || 1000
+      client
+    end
+    
+    def models_by_crc
+      @models_by_crc ||= begin
+        ThinkingSphinx.context.indexed_models.inject({}) do |hash, model|
+          hash[model.constantize.to_crc32] = model
+          Object.subclasses_of(model.constantize).each { |subclass|
+            hash[subclass.to_crc32] = subclass.name
+          }
+          hash
+        end
+      end
+    end
+    
+    private
+    
+    # Parse the config/sphinx.yml file - if it exists - then use the attribute
+    # accessors to set the appropriate values. Nothing too clever.
+    # 
+    def parse_config
+      path = "#{app_root}/config/sphinx.yml"
+      return unless File.exists?(path)
+      
+      conf = YAML::load(ERB.new(IO.read(path)).result)[environment]
+      
+      conf.each do |key,value|
+        self.send("#{key}=", value) if self.respond_to?("#{key}=")
+        
+        set_sphinx_setting self.source_options, key, value, SourceOptions
+        set_sphinx_setting self.index_options,  key, value, IndexOptions
+        set_sphinx_setting self.index_options,  key, value, CustomOptions
+        set_sphinx_setting @configuration.searchd, key, value
+        set_sphinx_setting @configuration.indexer, key, value
+      end unless conf.nil?
+      
+      self.bin_path += '/' unless self.bin_path.blank?
+      
+      if self.allow_star
+        self.index_options[:enable_star]    = true
+        self.index_options[:min_prefix_len] = 1
+      end
+    end
+    
+    def set_sphinx_setting(object, key, value, allowed = {})
+      if object.is_a?(Hash)
+        object[key.to_sym] = value if allowed.include?(key.to_s)
+      else
+        object.send("#{key}=", value) if object.respond_to?("#{key}")
+        send("#{key}=", value) if self.respond_to?("#{key}")
+      end
+    end
+  end
+end

data/lib/thinking_sphinx/context.rb

@@ -0,0 +1,68 @@
+class ThinkingSphinx::Context
+  attr_reader :indexed_models
+  
+  def initialize
+    @indexed_models = []
+  end
+  
+  def prepare
+    load_models
+    add_indexed_models
+  end
+  
+  def define_indexes
+    indexed_models.each { |model|
+      model.constantize.define_indexes
+    }
+  end
+  
+  def add_indexed_model(model)
+    model = model.name if model.is_a?(Class)
+    
+    indexed_models << model
+    indexed_models.uniq!
+    indexed_models.sort!
+  end
+  
+  def superclass_indexed_models
+    klasses = indexed_models.collect { |name| name.constantize }
+    klasses.reject { |klass|
+      klass.superclass.ancestors.any? { |ancestor| klasses.include?(ancestor) }
+    }.collect { |klass| klass.name }
+  end
+  
+  private
+  
+  def add_indexed_models
+    Object.subclasses_of(ActiveRecord::Base).each do |klass|
+      add_indexed_model klass if klass.has_sphinx_indexes?
+    end
+  end
+  
+  # Make sure all models are loaded - without reloading any that
+  # ActiveRecord::Base is already aware of (otherwise we start to hit some
+  # messy dependencies issues).
+  #
+  def load_models
+    ThinkingSphinx::Configuration.instance.model_directories.each do |base|
+      Dir["#{base}**/*.rb"].each do |file|
+        model_name = file.gsub(/^#{base}([\w_\/\\]+)\.rb/, '\1')
+      
+        next if model_name.nil?
+        next if ::ActiveRecord::Base.send(:subclasses).detect { |model|
+          model.name == model_name
+        }
+      
+        begin
+          model_name.camelize.constantize
+        rescue LoadError
+          model_name.gsub!(/.*[\/\\]/, '').nil? ? next : retry
+        rescue NameError
+          next
+        rescue StandardError
+          STDERR.puts "Warning: Error loading #{file}"
+        end
+      end
+    end
+  end
+end

data/lib/thinking_sphinx/core/array.rb

@@ -0,0 +1,7 @@
+module SearchAsArray
+  def ===(object)
+    object.is_a?(ThinkingSphinx::Search) || super
+  end
+end
+
+Array.extend SearchAsArray

data/lib/thinking_sphinx/core/string.rb

@@ -0,0 +1,15 @@
+require 'zlib'
+
+module ThinkingSphinx
+  module Core
+    module String
+      def to_crc32
+        Zlib.crc32 self
+      end
+    end
+  end
+end
+
+class String
+  include ThinkingSphinx::Core::String
+end

data/lib/thinking_sphinx/deltas/default_delta.rb

@@ -0,0 +1,62 @@
+module ThinkingSphinx
+  module Deltas
+    class DefaultDelta
+      attr_accessor :column
+      
+      def initialize(index, options)
+        @index  = index
+        @column = options.delete(:delta_column) || :delta
+      end
+      
+      def index(model, instance = nil)
+        return true unless ThinkingSphinx.updates_enabled? &&
+          ThinkingSphinx.deltas_enabled?
+        return true if instance && !toggled(instance)
+        
+        update_delta_indexes model
+        delete_from_core     model, instance if instance
+        
+        true
+      end
+      
+      def toggle(instance)
+        instance.delta = true
+      end
+      
+      def toggled(instance)
+        instance.delta
+      end
+      
+      def reset_query(model)
+        "UPDATE #{model.quoted_table_name} SET " +
+        "#{model.connection.quote_column_name(@column.to_s)} = #{adapter.boolean(false)} " +
+        "WHERE #{model.connection.quote_column_name(@column.to_s)} = #{adapter.boolean(true)}"
+      end
+      
+      def clause(model, toggled)
+        "#{model.quoted_table_name}.#{model.connection.quote_column_name(@column.to_s)}" +
+        " = #{adapter.boolean(toggled)}"
+      end
+      
+      private
+      
+      def update_delta_indexes(model)
+        config = ThinkingSphinx::Configuration.instance
+        rotate = ThinkingSphinx.sphinx_running? ? "--rotate" : ""
+        
+        output = `#{config.bin_path}#{config.indexer_binary_name} --config #{config.config_file} #{rotate} #{model.delta_index_names.join(' ')}`
+        puts(output) unless ThinkingSphinx.suppress_delta_output?
+      end
+      
+      def delete_from_core(model, instance)
+        model.core_index_names.each do |index_name|
+          model.delete_in_index index_name, instance.sphinx_document_id
+        end
+      end
+      
+      def adapter
+        @adapter = @index.model.sphinx_database_adapter
+      end
+    end
+  end
+end

data/lib/thinking_sphinx/deltas.rb

@@ -0,0 +1,28 @@
+require 'thinking_sphinx/deltas/default_delta'
+
+module ThinkingSphinx
+  module Deltas
+    def self.parse(index)
+      delta_option = index.local_options.delete(:delta)
+      case delta_option
+      when TrueClass, :default
+        DefaultDelta.new index, index.local_options
+      when :delayed
+        DelayedDelta.new index, index.local_options
+      when :datetime
+        DatetimeDelta.new index, index.local_options
+      when FalseClass, nil
+        nil
+      else
+        if delta_option.is_a?(String)
+          delta_option = Kernel.const_get(delta_option)
+        end
+        if delta_option.ancestors.include?(ThinkingSphinx::Deltas::DefaultDelta)
+          delta_option.new index, index.local_options
+        else
+          raise "Unknown delta type"
+        end
+      end
+    end
+  end
+end

data/lib/thinking_sphinx/deploy/capistrano.rb

@@ -0,0 +1,100 @@
+Capistrano::Configuration.instance(:must_exist).load do
+  namespace :thinking_sphinx do
+    namespace :install do
+      desc <<-DESC
+        Install Sphinx by source
+        
+        If Postgres is available, Sphinx will use it.
+        
+        If the variable :thinking_sphinx_configure_args is set, it will
+        be passed to the Sphinx configure script. You can use this to
+        install Sphinx in a non-standard location:
+        
+          set :thinking_sphinx_configure_args, "--prefix=$HOME/software"
+DESC
+
+      task :sphinx do
+        with_postgres = false
+        begin
+          run "which pg_config" do |channel, stream, data|
+            with_postgres = !(data.nil? || data == "")
+          end
+        rescue Capistrano::CommandError => e
+          puts "Continuing despite error: #{e.message}"
+        end
+      
+        args = []
+        if with_postgres
+          run "pg_config --pkgincludedir" do |channel, stream, data|
+            args << "--with-pgsql=#{data}"
+          end
+        end
+        args << fetch(:thinking_sphinx_configure_args, '')
+        
+        commands = <<-CMD
+        wget -q http://www.sphinxsearch.com/downloads/sphinx-0.9.8.1.tar.gz >> sphinx.log
+        tar xzvf sphinx-0.9.8.1.tar.gz
+        cd sphinx-0.9.8.1
+        ./configure #{args.join(" ")}
+        make
+        #{try_sudo} make install
+        rm -rf sphinx-0.9.8.1 sphinx-0.9.8.1.tar.gz
+        CMD
+        run commands.split(/\n\s+/).join(" && ")
+      end
+    
+      desc "Install Thinking Sphinx as a gem from GitHub"
+      task :ts do
+        run "#{try_sudo} gem install thinking-sphinx --source http://gemcutter.org"
+      end
+    end
+  
+    desc "Generate the Sphinx configuration file"
+    task :configure do
+      rake "thinking_sphinx:configure"
+    end
+  
+    desc "Index data"
+    task :index do
+      rake "thinking_sphinx:index"
+    end
+  
+    desc "Start the Sphinx daemon"
+    task :start do
+      configure
+      rake "thinking_sphinx:start"
+    end
+  
+    desc "Stop the Sphinx daemon"
+    task :stop do
+      configure
+      rake "thinking_sphinx:stop"
+    end
+  
+    desc "Stop and then start the Sphinx daemon"
+    task :restart do
+      stop
+      start
+    end
+  
+    desc "Stop, re-index and then start the Sphinx daemon"
+    task :rebuild do
+      stop
+      index
+      start
+    end
+  
+    desc "Add the shared folder for sphinx files for the production environment"
+    task :shared_sphinx_folder, :roles => :web do
+      run "mkdir -p #{shared_path}/db/sphinx/production"
+    end
+
+    def rake(*tasks)
+      rails_env = fetch(:rails_env, "production")
+      rake = fetch(:rake, "rake")
+      tasks.each do |t|
+        run "if [ -d #{release_path} ]; then cd #{release_path}; else cd #{current_path}; fi; #{rake} RAILS_ENV=#{rails_env} #{t}"
+      end
+    end
+  end
+end

data/lib/thinking_sphinx/excerpter.rb

@@ -0,0 +1,22 @@
+module ThinkingSphinx
+  class Excerpter
+    CoreMethods = %w( kind_of? object_id respond_to? should should_not stub! )
+    # Hide most methods, to allow them to be passed through to the instance.
+    instance_methods.select { |method|
+      method.to_s[/^__/].nil? && !CoreMethods.include?(method.to_s)
+    }.each { |method|
+      undef_method method
+    }
+    
+    def initialize(search, instance)
+      @search   = search
+      @instance = instance
+    end
+    
+    def method_missing(method, *args, &block)
+      string = CGI::escapeHTML @instance.send(method, *args, &block).to_s
+      
+      @search.excerpt_for(string, @instance.class)
+    end
+  end
+end