UnderpantsGnome-sunspot 0.9.1.1

data/lib/sunspot/dsl/search.rb

@@ -0,0 +1,30 @@
+module Sunspot
+  module DSL
+    # 
+    # This top-level DSL class is the context in which the block passed to
+    # Sunspot.query. See Sunspot::DSL::Query, Sunspot::DSL::FieldQuery, and
+    # Sunspot::DSL::Scope for the full API presented.
+    #
+    class Search < Query
+      def initialize(search) #:nodoc:
+        @search = search
+        @query = search.query
+      end
+
+      # 
+      # Retrieve the data accessor used to load instances of the given class
+      # out of persistent storage. Data accessors are free to implement any
+      # extra methods that may be useful in this context.
+      #
+      # ==== Example
+      #
+      #   Sunspot.search Post do
+      #     data_acccessor_for(Post).includes = [:blog, :comments]
+      #   end
+      #
+      def data_accessor_for(clazz)
+        @search.data_accessor_for(clazz)
+      end
+    end
+  end
+end

data/lib/sunspot/facet.rb

@@ -0,0 +1,51 @@
+require 'enumerator'
+
+module Sunspot
+  #
+  # The facet class encapsulates the information returned by Solr for a
+  # field facet request.
+  #
+  # See http://wiki.apache.org/solr/SolrFacetingOverview for more information
+  # on Solr's faceting capabilities.
+  #
+  class Facet
+    attr_reader :field
+
+    def initialize(facet_values, field) #:nodoc:
+      @facet_values, @field = facet_values, field
+    end
+
+    # The name of the field that contains this facet's values
+    #
+    # ==== Returns
+    #
+    # Symbol:: The field name
+    # 
+    def field_name
+      @field.name
+    end
+
+    # The rows returned for this facet.
+    #
+    # ==== Returns
+    #
+    # Array:: Collection of FacetRow objects, in the order returned by Solr
+    # 
+    def rows
+      @rows ||=
+        begin
+          rows = []
+          @facet_values.each_slice(2) do |pair|
+            rows << new_row(pair)
+          end
+          rows
+        end
+    end
+
+    private
+
+    def new_row(pair)
+      FacetRow.new(pair, self)
+    end
+  end
+end

data/lib/sunspot/facet_row.rb

@@ -0,0 +1,34 @@
+module Sunspot
+  # This class encapsulates a facet row (value) for a facet.
+  class FacetRow
+    def initialize(pair, facet) #:nodoc:
+      @pair, @facet = pair, facet
+    end
+
+    # The value associated with the facet. This will be cast according to the
+    # field's type; so, for an integer field, this method will return an
+    # integer, etc.
+    #
+    # Note that <strong>+Time+ fields will always return facet values in
+    # UTC</strong>.
+    #
+    # ==== Returns
+    #
+    # Object:: The value associated with the row, cast to the appropriate type
+    #
+    def value
+      @value ||= @facet.field.cast(@pair[0])
+    end
+
+    # The number of documents matching the search parameters that have this
+    # value in the facet's field.
+    #
+    # ==== Returns
+    #
+    # Integer:: Document count for this value
+    #
+    def count
+      @count ||= @pair[1]
+    end
+  end
+end

data/lib/sunspot/field.rb

@@ -0,0 +1,157 @@
+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_accessor :attributes
+
+    # 
+    #
+    def initialize(name, type) #:nodoc
+      @name, @type = name.to_sym, type
+      @attributes = {}
+    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
+
+    # 
+    # Name with which this field is indexed internally. Based on public name and
+    # type.
+    #
+    # ==== Returns
+    #
+    # String:: Internal name of the field
+    #
+    def indexed_name
+      @type.indexed_name(@name)
+    end
+
+    # 
+    # Whether this field accepts multiple values.
+    #
+    # ==== Returns
+    #
+    # Boolean:: True if this field accepts multiple values.
+    #
+    def multiple?
+      !!@multiple
+    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:
+    def initialize(name, options = {})
+      super(name, Type::TextType)
+      if options.has_key?(:boost)
+        @attributes[:boost] = options.delete(:boost)
+      end
+      @multiple = true
+      raise ArgumentError, "Unknown field option #{options.keys.first.inspect} provided for field #{name.inspect}" unless options.empty?
+    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 = {})
+      super(name, type)
+      @multiple = !!options.delete(:multiple)
+      @reference =
+        if (reference = options.delete(:references)).respond_to?(:name)
+          reference.name
+        elsif reference.respond_to?(:to_sym)
+          reference.to_sym
+        end
+      @stored = !!options.delete(:stored)
+      raise ArgumentError, "Unknown field option #{options.keys.first.inspect} provided for field #{name.inspect}" unless options.empty?
+    end
+
+    # The name of the field as it is indexed in Solr. The indexed name
+    # contains a suffix that contains information about the type as well as
+    # whether the field allows multiple values for a document.
+    #
+    # ==== Returns
+    #
+    # String:: The field's indexed name
+    #
+    def indexed_name
+      "#{super}#{'m' if @multiple}#{'s' if @stored}"
+    end
+  end
+
+  # 
+  # RandomField instances are used for random sorting.
+  #
+  class RandomField #:nodoc:
+    # 
+    # Never multiple, but this has to return false so Sunspot doesn't barf
+    # when you try to order by it.
+    #
+    def multiple?
+      false
+    end
+
+    # 
+    # Solr uses the dynamic field name as a seed for random, so we randomize the
+    # field name accordingly.
+    #
+    # #XXX I think it's bad to use a random number as a seed. Would it be
+    # better to pass in the current timestamp or some such thing?
+    #
+    def indexed_name
+      "random_#{rand(1<<16)}"
+    end
+  end
+end

data/lib/sunspot/field_factory.rb

@@ -0,0 +1,126 @@
+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 RSolr 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 == 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 RSolr document for indexing.
+      #
+      def populate_document(document, model) #:nodoc:
+        unless (value = @data_extractor.value_for(model)).nil?
+          for scalar_value in Array(@field.to_indexed(value))
+            document.add_field(
+              @field.indexed_name.to_sym,
+              scalar_value, @field.attributes
+            )
+          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)
+            for scalar_value in Array(field_instance.to_indexed(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 +1,127 @@
+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 = Array(model).map { |m| prepare(m) }
+      if @batch.nil?
+        add_documents(documents)
+      else
+        @batch.concat(documents)
+      end
+    end
+
+    #
+    # Remove the given model from the Solr index
+    #
+    def remove(model)
+      @connection.delete_by_id(Adapters::InstanceAdapter.adapt(model).index_id)
+    end
+
+    def remove_by_id(class_name, id)
+      @connection.delete_by_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)
+      @connection.delete_by_query("type:#{escape(clazz.name)}")
+    end
+
+    def start_batch
+      @batch = []
+    end
+
+    def flush_batch
+      add_documents(@batch)
+      @batch = nil
+    end
+
+    private
+
+    #
+    # 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
+      for field_factory in setup.all_field_factories
+        field_factory.populate_document(document, model)
+      end
+      document
+    end
+
+    def add_documents(documents)
+      begin
+        @connection.add(documents)
+      rescue Exception => e
+        puts e
+      end
+    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::Message::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
+
+
+    class <<self
+      #
+      # Delete all documents from the Solr index
+      #
+      # ==== Parameters
+      #
+      # connection<Solr::Connection>::
+      #   connection to which to send the delete request
+      def remove_all(connection)
+        connection.delete_by_query("type:[* TO *]")
+      end
+    end
+  end
+end