gojee-sunspot 2.0.3 → 2.0.4

data/lib/sunspot/dsl/standard_query.rb

@@ -0,0 +1,123 @@
+module Sunspot
+  module DSL #:nodoc:
+    #
+    # This class presents a DSL for constructing queries using the
+    # Sunspot.search method. Methods of this class are available inside the
+    # search block. Much of the DSL's functionality is implemented by this
+    # class's superclasses, Sunspot::DSL::FieldQuery and Sunspot::DSL::Scope
+    #
+    # See Sunspot.search for usage examples
+    #
+    class StandardQuery < FieldQuery
+      include Paginatable, Adjustable
+
+      # Specify a phrase that should be searched as fulltext. Only +text+
+      # fields are searched - see DSL::Fields.text
+      #
+      # Keyword search is executed using Solr's dismax handler, which strikes
+      # a good balance between powerful and foolproof. In particular,
+      # well-matched quotation marks can be used to group phrases, and the
+      # + and - modifiers work as expected. All other special Solr boolean
+      # syntax is escaped, and mismatched quotes are ignored entirely.
+      #
+      # This method can optionally take a block, which is evaluated by the
+      # Fulltext DSL class, and exposes several powerful dismax features.
+      #
+      # ==== Parameters
+      #
+      # keywords<String>:: phrase to perform fulltext search on
+      #
+      # ==== Options
+      #
+      # :fields<Array>::
+      #   List of fields that should be searched for keywords. Defaults to all
+      #   fields configured for the types under search.
+      # :highlight<Boolean,Array>::
+      #   If true, perform keyword highlighting on all searched fields. If an
+      #   array of field names, perform highlighting on the specified fields.
+      #   Note that for highlighting to work, the desired fields have to be set
+      #   up with :stored => true.
+      #   This can also be called from within the fulltext block.
+      # :minimum_match<Integer>::
+      #   The minimum number of search terms that a result must match. By
+      #   default, all search terms must match; if the number of search terms
+      #   is less than this number, the default behavior applies.
+      # :tie<Float>::
+      #   A tiebreaker coefficient for scores derived from subqueries that are
+      #   lower-scoring than the maximum score subquery. Typically a near-zero
+      #   value is useful. See
+      #   http://wiki.apache.org/solr/DisMaxRequestHandler#tie_.28Tie_breaker.29
+      #   for more information.
+      # :query_phrase_slop<Integer>::
+      #   The number of words that can appear between the words in a
+      #   user-entered phrase (i.e., keywords in quotes) and still match. For
+      #   instance, in a search for "\"great pizza\"" with a phrase slop of 1,
+      #   "great pizza" and "great big pizza" will match, but "great monster of
+      #   a pizza" will not. Default behavior is a query phrase slop of zero.
+      #
+      def fulltext(keywords, options = {}, &block)
+        if keywords && !(keywords.to_s =~ /^\s*$/)
+          fulltext_query = @query.add_fulltext(keywords)
+          if field_names = options.delete(:fields)
+            Util.Array(field_names).each do |field_name|
+              @setup.text_fields(field_name).each do |field|
+                fulltext_query.add_fulltext_field(field, field.default_boost)
+              end
+            end
+          end
+          if minimum_match = options.delete(:minimum_match)
+            fulltext_query.minimum_match = minimum_match.to_i
+          end
+          if tie = options.delete(:tie)
+            fulltext_query.tie = tie.to_f
+          end
+          if query_phrase_slop = options.delete(:query_phrase_slop)
+            fulltext_query.query_phrase_slop = query_phrase_slop.to_i
+          end
+          if highlight_field_names = options.delete(:highlight)
+            if highlight_field_names == true
+              fulltext_query.add_highlight
+            else
+              highlight_fields = []
+              Util.Array(highlight_field_names).each do |field_name|
+                highlight_fields.concat(@setup.text_fields(field_name))
+              end
+              fulltext_query.add_highlight(highlight_fields)
+            end
+          end
+          if block && fulltext_query
+            fulltext_dsl = Fulltext.new(fulltext_query, @setup)
+            Util.instance_eval_or_call(
+              fulltext_dsl,
+              &block
+            )
+          end
+          if !field_names && (!fulltext_dsl || !fulltext_dsl.fields_added?)
+            @setup.all_text_fields.each do |field|
+              unless fulltext_query.has_fulltext_field?(field)
+                unless fulltext_dsl && fulltext_dsl.exclude_fields.include?(field.name)
+                  fulltext_query.add_fulltext_field(field, field.default_boost)
+                end
+              end
+            end
+          end
+        end
+      end
+      alias_method :keywords, :fulltext
+
+      def with(*args)
+        case args.first
+        when String, Symbol
+          field_name = args[0]
+          value = args.length > 1 ? args[1] : Scope::NONE
+          if value == Scope::NONE
+            return DSL::RestrictionWithNear.new(@setup.field(field_name.to_sym), @scope, @query, false)
+          end
+        end
+
+        # else
+        super
+      end
+    end
+  end
+end

data/lib/sunspot/dsl.rb

@@ -0,0 +1,5 @@
+%w(fields scope paginatable adjustable field_query standard_query query_facet
+   functional fulltext restriction restriction_with_near search
+   more_like_this_query function field_group).each do |file|
+  require File.join(File.dirname(__FILE__), 'dsl', file)
+end

data/lib/sunspot/field.rb

@@ -0,0 +1,193 @@
+module Sunspot
+  class Field #:nodoc:
+    attr_accessor :name # The public-facing name of the field
+    attr_accessor :type # The Type of the field
+    attr_accessor :reference # Model class that the value of this field refers to
+    attr_reader :boost
+    attr_reader :indexed_name # Name with which this field is indexed internally. Based on public name and type or the +:as+ option.
+
+    #
+    #
+    def initialize(name, type, options = {}) #:nodoc
+      @name, @type = name.to_sym, type
+      @stored = !!options.delete(:stored)
+      @more_like_this = !!options.delete(:more_like_this)
+      set_indexed_name(options)
+      raise ArgumentError, "Field of type #{type} cannot be used for more_like_this" unless type.accepts_more_like_this? or !@more_like_this
+    end
+
+    # Convert a value to its representation for Solr indexing. This delegates
+    # to the #to_indexed method on the field's type.
+    #
+    # ==== Parameters
+    #
+    # value<Object>:: Value to convert to Solr representation
+    #
+    # ==== Returns
+    #
+    # String:: Solr representation of the object
+    #
+    # ==== Raises
+    #
+    # ArgumentError::
+    #   the value is an array, but this field does not allow multiple values
+    #
+    def to_indexed(value)
+      if value.is_a? Array
+        if @multiple
+          value.map { |val| to_indexed(val) }
+        else
+          raise ArgumentError, "#{name} is not a multiple-value field, so it cannot index values #{value.inspect}"
+        end
+      else
+        @type.to_indexed(value)
+      end
+    end
+
+    # Cast the value into the appropriate Ruby class for the field's type
+    #
+    # ==== Parameters
+    #
+    # value<String>:: Solr's representation of the value
+    #
+    # ==== Returns
+    #
+    # Object:: The cast value
+    #
+    def cast(value)
+      @type.cast(value)
+    end
+
+    #
+    # Whether this field accepts multiple values.
+    #
+    # ==== Returns
+    #
+    # Boolean:: True if this field accepts multiple values.
+    #
+    def multiple?
+      !!@multiple
+    end
+
+    #
+    # Whether this field can be used for more_like_this queries.
+    # If true, the field is configured to store termVectors.
+    #
+    # ==== Returns
+    #
+    # Boolean:: True if this field can be used for more_like_this queries.
+    #
+    def more_like_this?
+      !!@more_like_this
+    end
+
+    def hash
+      indexed_name.hash
+    end
+
+    def eql?(field)
+      indexed_name == field.indexed_name
+    end
+    alias_method :==, :eql?
+
+    private
+
+    #
+    # Determine the indexed name. If the :as option is given use that, otherwise
+    # create the value based on the indexed_name of the type with additional
+    # suffixes for multiple, stored, and more_like_this.
+    #
+    # ==== Returns
+    #
+    # String: The field's indexed name
+    #
+    def set_indexed_name(options)
+      @indexed_name =
+        if options[:as]
+          options.delete(:as)
+        else
+          "#{@type.indexed_name(@name).to_s}#{'m' if @multiple }#{'s' if @stored}#{'v' if more_like_this?}"
+        end
+    end
+
+  end
+
+  #
+  # FulltextField instances represent fields that are indexed as fulltext.
+  # These fields are tokenized in the index, and can have boost applied to
+  # them. They also always allow multiple values (since the only downside of
+  # allowing multiple values is that it prevents the field from being sortable,
+  # and sorting on tokenized fields is nonsensical anyway, there is no reason
+  # to do otherwise). FulltextField instances always have the type TextType.
+  #
+  class FulltextField < Field #:nodoc:
+    attr_reader :default_boost
+
+    def initialize(name, options = {})
+      super(name, Type::TextType.instance, options)
+      @multiple = true
+      @boost = options.delete(:boost)
+      @default_boost = options.delete(:default_boost)
+      raise ArgumentError, "Unknown field option #{options.keys.first.inspect} provided for field #{name.inspect}" unless options.empty?
+    end
+
+    def indexed_name
+      "#{super}"
+    end
+  end
+
+  #
+  # AttributeField instances encapsulate non-tokenized attribute data.
+  # AttributeFields can have any type except TextType, and can also have
+  # a reference (for instantiated facets), optionally allow multiple values
+  # (false by default), and can store their values (false by default). All
+  # scoping, sorting, and faceting is done with attribute fields.
+  #
+  class AttributeField < Field #:nodoc:
+    def initialize(name, type, options = {})
+      @multiple = !!options.delete(:multiple)
+      super(name, type, options)
+      @reference =
+        if (reference = options.delete(:references)).respond_to?(:name)
+          reference.name
+        elsif reference.respond_to?(:to_sym)
+          reference.to_sym
+        end
+      raise ArgumentError, "Unknown field option #{options.keys.first.inspect} provided for field #{name.inspect}" unless options.empty?
+    end
+
+  end
+
+  class TypeField #:nodoc:
+    class <<self
+      def instance
+        @instance ||= new
+      end
+    end
+
+    def indexed_name
+      'type'
+    end
+
+    def to_indexed(clazz)
+      clazz.name
+    end
+  end
+
+  class IdField #:nodoc:
+    class <<self
+      def instance
+        @instance ||= new
+      end
+    end
+
+    def indexed_name
+      'id'
+    end
+
+    def to_indexed(id)
+      id.to_s
+    end
+  end
+end
+

data/lib/sunspot/field_factory.rb

@@ -0,0 +1,129 @@
+module Sunspot
+  # 
+  # The FieldFactory module contains classes for generating fields. FieldFactory
+  # implementation classes should implement a #build method, although the arity
+  # of the method depends on the type of factory. They also must implement a
+  # #populate_document method, which extracts field data from a given model and
+  # adds it into the Solr document for indexing.
+  #
+  module FieldFactory #:nodoc:all
+    # 
+    # Base class for field factories.
+    #
+    class Abstract
+      attr_reader :name
+
+      def initialize(name, options = {}, &block)
+        @name = name.to_sym
+        @data_extractor =
+          if block
+            DataExtractor::BlockExtractor.new(&block)
+          else
+            DataExtractor::AttributeExtractor.new(options.delete(:using) || name)
+          end
+      end
+    end
+
+    # 
+    # A StaticFieldFactory generates normal static fields. Each factory instance
+    # contains an eager-initialized field instance, which is returned by the
+    # #build method.
+    #
+    class Static < Abstract
+      def initialize(name, type, options = {}, &block)
+        super(name, options, &block)
+        unless name.to_s =~ /^\w+$/
+          raise ArgumentError, "Invalid field name #{name}: only letters, numbers, and underscores are allowed."
+        end
+        @field =
+          if type.is_a?(Type::TextType)
+            FulltextField.new(name, options)
+          else
+            AttributeField.new(name, type, options)
+          end
+      end
+
+      # 
+      # Return the field instance built by this factory
+      #
+      def build
+        @field
+      end
+
+      # 
+      # Extract the encapsulated field's data from the given model and add it
+      # into the Solr document for indexing.
+      #
+      def populate_document(document, model) #:nodoc:
+        unless (value = @data_extractor.value_for(model)).nil?
+          Util.Array(@field.to_indexed(value)).each do |scalar_value|
+            options = {}
+            options[:boost] = @field.boost if @field.boost
+            document.add_field(
+              @field.indexed_name.to_sym,
+              scalar_value,
+              options
+            )
+          end
+        end
+      end
+
+      # 
+      # A unique signature identifying this field by name and type.
+      #
+      def signature
+        [@field.name, @field.type]
+      end
+    end
+
+    # 
+    # DynamicFieldFactories create dynamic field instances based on dynamic
+    # configuration.
+    #
+    class Dynamic < Abstract
+      attr_accessor :name, :type
+
+      def initialize(name, type, options = {}, &block)
+        super(name, options, &block)
+        @type, @options = type, options
+      end
+
+      #
+      # Build a field based on the dynamic name given.
+      #
+      def build(dynamic_name)
+        AttributeField.new("#{@name}:#{dynamic_name}", @type, @options.dup)
+      end
+      # 
+      # This alias allows a DynamicFieldFactory to be used in place of a Setup
+      # or CompositeSetup instance by query components.
+      #
+      alias_method :field, :build
+
+      # 
+      # Generate dynamic fields based on hash returned by data accessor and
+      # add the field data to the document.
+      #
+      def populate_document(document, model)
+        if values = @data_extractor.value_for(model)
+          values.each_pair do |dynamic_name, value|
+            field_instance = build(dynamic_name)
+            Util.Array(field_instance.to_indexed(value)).each do |scalar_value|
+              document.add_field(
+                field_instance.indexed_name.to_sym,
+                scalar_value
+              )
+            end
+          end
+        end
+      end
+
+      # 
+      # Unique signature identifying this dynamic field based on name and type
+      #
+      def signature
+        [@name, @type]
+      end
+    end
+  end
+end

data/lib/sunspot/indexer.rb

@@ -0,0 +1,136 @@
+require 'sunspot/batcher'
+
+module Sunspot
+  # 
+  # This class presents a service for adding, updating, and removing data
+  # from the Solr index. An Indexer instance is associated with a particular
+  # setup, and thus is capable of indexing instances of a certain class (and its
+  # subclasses).
+  #
+  class Indexer #:nodoc:
+    include RSolr::Char
+
+    def initialize(connection)
+      @connection = connection
+    end
+
+    # 
+    # Construct a representation of the model for indexing and send it to the
+    # connection for indexing
+    #
+    # ==== Parameters
+    #
+    # model<Object>:: the model to index
+    #
+    def add(model)
+      documents = Util.Array(model).map { |m| prepare(m) }
+      if batcher.batching?
+        batcher.concat(documents)
+      else
+        add_documents(documents)
+      end
+    end
+
+    # 
+    # Remove the given model from the Solr index
+    #
+    def remove(*models)
+      @connection.delete_by_id(
+        models.map { |model| Adapters::InstanceAdapter.adapt(model).index_id }
+      )
+    end
+
+    # 
+    # Remove the model from the Solr index by specifying the class and ID
+    #
+    def remove_by_id(class_name, *ids)
+      @connection.delete_by_id(
+        ids.map { |id| Adapters::InstanceAdapter.index_id_for(class_name, id) }
+      )
+    end
+
+    # 
+    # Delete all documents of the class indexed by this indexer from Solr.
+    #
+    def remove_all(clazz = nil)
+      if clazz
+        @connection.delete_by_query("type:#{escape(clazz.name)}")
+      else
+        @connection.delete_by_query("*:*")
+      end
+    end
+
+    # 
+    # Remove all documents that match the scope given in the Query
+    #
+    def remove_by_scope(scope)
+      @connection.delete_by_query(scope.to_boolean_phrase)
+    end
+
+    # 
+    # Start batch processing
+    #
+    def start_batch
+      batcher.start_new
+    end
+
+    #
+    # Write batch out to Solr and clear it
+    #
+    def flush_batch
+      add_documents(batcher.end_current)
+    end
+
+    private
+
+    def batcher
+      @batcher ||= Batcher.new
+    end
+
+    # 
+    # Convert documents into hash of indexed properties
+    #
+    def prepare(model)
+      document = document_for(model)
+      setup = setup_for(model)
+      if boost = setup.document_boost_for(model)
+        document.attrs[:boost] = boost
+      end
+      setup.all_field_factories.each do |field_factory|
+        field_factory.populate_document(document, model)
+      end
+      document
+    end
+
+    def add_documents(documents)
+      @connection.add(documents)
+    end
+
+    # 
+    # All indexed documents index and store the +id+ and +type+ fields.
+    # This method constructs the document hash containing those key-value
+    # pairs.
+    #
+    def document_for(model)
+      RSolr::Xml::Document.new(
+        :id => Adapters::InstanceAdapter.adapt(model).index_id,
+        :type => Util.superclasses_for(model.class).map { |clazz| clazz.name }
+      )
+    end
+
+    # 
+    # Get the Setup object for the given object's class.
+    #
+    # ==== Parameters
+    #
+    # object<Object>:: The object whose setup is to be retrieved
+    #
+    # ==== Returns
+    #
+    # Sunspot::Setup:: The setup for the object's class
+    #
+    def setup_for(object)
+      Setup.for(object.class) || raise(NoSetupError, "Sunspot is not configured for #{object.class.inspect}")
+    end
+  end
+end

data/lib/sunspot/query/abstract_field_facet.rb

@@ -0,0 +1,52 @@
+module Sunspot
+  module Query
+    class AbstractFieldFacet
+      include RSolr::Char
+
+      def initialize(field, options)
+        @field, @options = field, options
+      end
+
+      def to_params
+        params = {
+          :facet => 'true',
+        }
+        case @options[:sort]
+        when :count
+          params[qualified_param('sort')] = 'true'
+        when :index
+          params[qualified_param('sort')] = 'false'
+        when nil
+        else
+          raise(
+            ArgumentError,
+            "#{@options[:sort].inspect} is not an allowed value for :sort. Allowed options are :count and :index"
+          )
+        end
+        if @options[:limit]
+          params[qualified_param('limit')] = @options[:limit].to_i
+        end
+        if @options[:prefix]
+          params[qualified_param('prefix')] = escape(@options[:prefix].to_s)
+        end
+        params[qualified_param('mincount')] = 
+          case
+          when @options[:minimum_count] then @options[:minimum_count].to_i
+          when @options[:zeros] then 0
+          else 1
+          end
+        params
+      end
+
+      private
+
+      def qualified_param(param)
+        :"f.#{key}.facet.#{param}"
+      end
+      
+      def key
+        @key ||= @options[:name] || @field.indexed_name
+      end
+    end
+  end
+end

data/lib/sunspot/query/bbox.rb

@@ -0,0 +1,15 @@
+module Sunspot
+  module Query
+    class Bbox
+      def initialize(field, first_corner, second_corner)
+        @field, @first_corner, @second_corner = field, first_corner, second_corner
+      end
+
+      def to_params
+        filter = "#{@field.indexed_name}:[#{@first_corner.join(",")} TO #{@second_corner.join(",")}]"
+
+        {:fq => filter}
+      end
+    end
+  end
+end

data/lib/sunspot/query/boost_query.rb

@@ -0,0 +1,24 @@
+module Sunspot
+  module Query
+    # 
+    # Representation of a BoostQuery, which allows the searcher to specify a
+    # scope for which matching documents should have an extra boost. This is
+    # essentially a conjunction, with an extra instance variable containing
+    # the boost that should be applied.
+    #
+    class BoostQuery < Connective::Conjunction #:nodoc:
+      def initialize(boost)
+        super(false)
+        @boost = boost
+      end
+
+      def to_boolean_phrase
+        if @boost.is_a?(FunctionQuery)
+          "#{@boost}"
+        else
+          "#{super}^#{@boost}"
+        end
+      end
+    end
+  end
+end