searchkick 2.3.2 → 5.2.1

data/lib/searchkick/relation_indexer.rb

@@ -0,0 +1,155 @@
+module Searchkick
+  class RelationIndexer
+    attr_reader :index
+
+    def initialize(index)
+      @index = index
+    end
+
+    def reindex(relation, mode:, method_name: nil, full: false, resume: false, scope: nil)
+      # apply scopes
+      if scope
+        relation = relation.send(scope)
+      elsif relation.respond_to?(:search_import)
+        relation = relation.search_import
+      end
+
+      # remove unneeded loading for async and queue
+      if mode == :async || mode == :queue
+        if relation.respond_to?(:primary_key)
+          relation = relation.except(:includes, :preload)
+          unless mode == :queue && relation.klass.method_defined?(:search_routing)
+            relation = relation.except(:select).select(relation.primary_key)
+          end
+        elsif relation.respond_to?(:only)
+          unless mode == :queue && relation.klass.method_defined?(:search_routing)
+            relation = relation.only(:_id)
+          end
+        end
+      end
+
+      if mode == :async && full
+        return full_reindex_async(relation)
+      end
+
+      relation = resume_relation(relation) if resume
+
+      reindex_options = {
+        mode: mode,
+        method_name: method_name,
+        full: full
+      }
+      record_indexer = RecordIndexer.new(index)
+
+      in_batches(relation) do |items|
+        record_indexer.reindex(items, **reindex_options)
+      end
+    end
+
+    def batches_left
+      Searchkick.with_redis { |r| r.call("SCARD", batches_key) }
+    end
+
+    def batch_completed(batch_id)
+      Searchkick.with_redis { |r| r.call("SREM", batches_key, [batch_id]) }
+    end
+
+    private
+
+    def resume_relation(relation)
+      if relation.respond_to?(:primary_key)
+        # use total docs instead of max id since there's not a great way
+        # to get the max _id without scripting since it's a string
+        where = relation.arel_table[relation.primary_key].gt(index.total_docs)
+        relation = relation.where(where)
+      else
+        raise Error, "Resume not supported for Mongoid"
+      end
+    end
+
+    def in_batches(relation)
+      if relation.respond_to?(:find_in_batches)
+        klass = relation.klass
+        # remove order to prevent possible warnings
+        relation.except(:order).find_in_batches(batch_size: batch_size) do |batch|
+          # prevent scope from affecting search_data as well as inline jobs
+          # Active Record runs relation calls in scoping block
+          # https://github.com/rails/rails/blob/main/activerecord/lib/active_record/relation/delegation.rb
+          # note: we could probably just call klass.current_scope = nil
+          # anywhere in reindex method (after initial all call),
+          # but this is more cautious
+          previous_scope = klass.current_scope(true)
+          if previous_scope
+            begin
+              klass.current_scope = nil
+              yield batch
+            ensure
+              klass.current_scope = previous_scope
+            end
+          else
+            yield batch
+          end
+        end
+      else
+        klass = relation.klass
+        each_batch(relation, batch_size: batch_size) do |batch|
+          # prevent scope from affecting search_data as well as inline jobs
+          # note: Model.with_scope doesn't always restore scope, so use custom logic
+          previous_scope = Mongoid::Threaded.current_scope(klass)
+          if previous_scope
+            begin
+              Mongoid::Threaded.set_current_scope(nil, klass)
+              yield batch
+            ensure
+              Mongoid::Threaded.set_current_scope(previous_scope, klass)
+            end
+          else
+            yield batch
+          end
+        end
+      end
+    end
+
+    def each_batch(relation, batch_size:)
+      # https://github.com/karmi/tire/blob/master/lib/tire/model/import.rb
+      # use cursor for Mongoid
+      items = []
+      relation.all.each do |item|
+        items << item
+        if items.length == batch_size
+          yield items
+          items = []
+        end
+      end
+      yield items if items.any?
+    end
+
+    def batch_size
+      @batch_size ||= index.options[:batch_size] || 1000
+    end
+
+    def full_reindex_async(relation)
+      batch_id = 1
+      class_name = relation.searchkick_options[:class_name]
+
+      in_batches(relation) do |items|
+        batch_job(class_name, batch_id, items.map(&:id))
+        batch_id += 1
+      end
+    end
+
+    def batch_job(class_name, batch_id, record_ids)
+      Searchkick.with_redis { |r| r.call("SADD", batches_key, [batch_id]) }
+      Searchkick::BulkReindexJob.perform_later(
+        class_name: class_name,
+        index_name: index.name,
+        batch_id: batch_id,
+        record_ids: record_ids.map { |v| v.instance_of?(Integer) ? v : v.to_s }
+      )
+    end
+
+    def batches_key
+      "searchkick:reindex:#{index.name}:batches"
+    end
+  end
+end

data/lib/searchkick/results.rb

@@ -1,10 +1,9 @@
-require "forwardable"
-
 module Searchkick
   class Results
     include Enumerable
     extend Forwardable
 
+    # TODO remove klass and options in 6.0
     attr_reader :klass, :response, :options
 
     def_delegators :results, :each, :any?, :empty?, :size, :length, :slice, :[], :to_ary
@@ -15,87 +14,33 @@ module Searchkick
       @options = options
     end
 
-    # experimental: may not make next release
-    def records
-      @records ||= results_query(klass, hits)
-    end
-
+    # TODO make private in 6.0
     def results
-      @results ||= begin
-        if options[:load]
-          # results can have different types
-          results = {}
-
-          hits.group_by { |hit, _| hit["_type"] }.each do |type, grouped_hits|
-            results[type] = results_query(type.camelize.constantize, grouped_hits).to_a.index_by { |r| r.id.to_s }
-          end
-
-          # sort
-          hits.map do |hit|
-            result = results[hit["_type"]][hit["_id"].to_s]
-            if result && !(options[:load].is_a?(Hash) && options[:load][:dumpable])
-              unless result.respond_to?(:search_hit)
-                result.define_singleton_method(:search_hit) do
-                  hit
-                end
-              end
-
-              if hit["highlight"] && !result.respond_to?(:search_highlights)
-                highlights = Hash[hit["highlight"].map { |k, v| [(options[:json] ? k : k.sub(/\.#{@options[:match_suffix]}\z/, "")).to_sym, v.first] }]
-                result.define_singleton_method(:search_highlights) do
-                  highlights
-                end
-              end
-            end
-            result
-          end.compact
-        else
-          hits.map do |hit|
-            result =
-              if hit["_source"]
-                hit.except("_source").merge(hit["_source"])
-              elsif hit["fields"]
-                hit.except("fields").merge(hit["fields"])
-              else
-                hit
-              end
+      @results ||= with_hit.map(&:first)
+    end
 
-            if hit["highlight"]
-              highlight = Hash[hit["highlight"].map { |k, v| [base_field(k), v.first] }]
-              options[:highlighted_fields].map { |k| base_field(k) }.each do |k|
-                result["highlighted_#{k}"] ||= (highlight[k] || result[k])
-              end
-            end
+    def with_hit
+      return enum_for(:with_hit) unless block_given?
 
-            result["id"] ||= result["_id"] # needed for legacy reasons
-            Hashie::Mash.new(result)
-          end
-        end
+      build_hits.each do |result|
+        yield result
       end
     end
 
+    def missing_records
+      @missing_records ||= with_hit_and_missing_records[1]
+    end
+
     def suggestions
       if response["suggest"]
         response["suggest"].values.flat_map { |v| v.first["options"] }.sort_by { |o| -o["score"] }.map { |o| o["text"] }.uniq
+      elsif options[:suggest] || options[:term] == "*" # TODO remove 2nd term
+        []
       else
         raise "Pass `suggest: true` to the search method for suggestions"
       end
     end
 
-    def each_with_hit(&block)
-      results.zip(hits).each(&block)
-    end
-
-    def with_details
-      each_with_hit.map do |model, hit|
-        details = {}
-        if hit["highlight"]
-          details[:highlight] = Hash[hit["highlight"].map { |k, v| [(options[:json] ? k : k.sub(/\.#{@options[:match_suffix]}\z/, "")).to_sym, v.first] }]
-        end
-        [model, details]
-      end
-    end
-
     def aggregations
       response["aggregations"]
     end
@@ -124,7 +69,11 @@ module Searchkick
     end
 
     def model_name
-      klass.model_name
+      if klass.nil?
+        ActiveModel::Name.new(self.class, nil, 'Result')
+      else
+        klass.model_name
+      end
     end
 
     def entry_name(options = {})
@@ -138,7 +87,13 @@ module Searchkick
     end
 
     def total_count
-      response["hits"]["total"]
+      if options[:total_entries]
+        options[:total_entries]
+      elsif response["hits"]["total"].is_a?(Hash)
+        response["hits"]["total"]["value"]
+      else
+        response["hits"]["total"]
+      end
     end
     alias_method :total_entries, :total_count
 
@@ -187,32 +142,188 @@ module Searchkick
     end
 
     def hits
-      @response["hits"]["hits"]
+      if error
+        raise Error, "Query error - use the error method to view it"
+      else
+        @response["hits"]["hits"]
+      end
+    end
+
+    def highlights(multiple: false)
+      hits.map do |hit|
+        hit_highlights(hit, multiple: multiple)
+      end
+    end
+
+    def with_highlights(multiple: false)
+      return enum_for(:with_highlights, multiple: multiple) unless block_given?
+
+      with_hit.each do |result, hit|
+        yield result, hit_highlights(hit, multiple: multiple)
+      end
+    end
+
+    def with_score
+      return enum_for(:with_score) unless block_given?
+
+      with_hit.each do |result, hit|
+        yield result, hit["_score"]
+      end
     end
 
     def misspellings?
       @options[:misspellings]
     end
 
+    def scroll_id
+      @response["_scroll_id"]
+    end
+
+    def scroll
+      raise Error, "Pass `scroll` option to the search method for scrolling" unless scroll_id
+
+      if block_given?
+        records = self
+        while records.any?
+          yield records
+          records = records.scroll
+        end
+
+        records.clear_scroll
+      else
+        begin
+          # TODO Active Support notifications for this scroll call
+          Results.new(@klass, Searchkick.client.scroll(scroll: options[:scroll], body: {scroll_id: scroll_id}), @options)
+        rescue => e
+          if Searchkick.not_found_error?(e) && e.message =~ /search_context_missing_exception/i
+            raise Error, "Scroll id has expired"
+          else
+            raise e
+          end
+        end
+      end
+    end
+
+    def clear_scroll
+      begin
+        # try to clear scroll
+        # not required as scroll will expire
+        # but there is a cost to open scrolls
+        Searchkick.client.clear_scroll(scroll_id: scroll_id)
+      rescue => e
+        raise e unless Searchkick.transport_error?(e)
+      end
+    end
+
     private
 
+    def with_hit_and_missing_records
+      @with_hit_and_missing_records ||= begin
+        missing_records = []
+
+        if options[:load]
+          grouped_hits = hits.group_by { |hit, _| hit["_index"] }
+
+          # determine models
+          index_models = {}
+          grouped_hits.each do |index, _|
+            models =
+              if @klass
+                [@klass]
+              else
+                index_alias = index.split("_")[0..-2].join("_")
+                Array((options[:index_mapping] || {})[index_alias])
+              end
+            raise Error, "Unknown model for index: #{index}. Pass the `models` option to the search method." unless models.any?
+            index_models[index] = models
+          end
+
+          # fetch results
+          results = {}
+          grouped_hits.each do |index, index_hits|
+            results[index] = {}
+            index_models[index].each do |model|
+              results[index].merge!(results_query(model, index_hits).to_a.index_by { |r| r.id.to_s })
+            end
+          end
+
+          # sort
+          results =
+            hits.map do |hit|
+              result = results[hit["_index"]][hit["_id"].to_s]
+              if result && !(options[:load].is_a?(Hash) && options[:load][:dumpable])
+                if (hit["highlight"] || options[:highlight]) && !result.respond_to?(:search_highlights)
+                  highlights = hit_highlights(hit)
+                  result.define_singleton_method(:search_highlights) do
+                    highlights
+                  end
+                end
+              end
+              [result, hit]
+            end.select do |result, hit|
+              unless result
+                models = index_models[hit["_index"]]
+                missing_records << {
+                  id: hit["_id"],
+                  # may be multiple models for inheritance with child models
+                  # not ideal to return different types
+                  # but this situation shouldn't be common
+                  model: models.size == 1 ? models.first : models
+                }
+              end
+              result
+            end
+        else
+          results =
+            hits.map do |hit|
+              result =
+                if hit["_source"]
+                  hit.except("_source").merge(hit["_source"])
+                elsif hit["fields"]
+                  hit.except("fields").merge(hit["fields"])
+                else
+                  hit
+                end
+
+              if hit["highlight"] || options[:highlight]
+                highlight = hit["highlight"].to_a.to_h { |k, v| [base_field(k), v.first] }
+                options[:highlighted_fields].map { |k| base_field(k) }.each do |k|
+                  result["highlighted_#{k}"] ||= (highlight[k] || result[k])
+                end
+              end
+
+              result["id"] ||= result["_id"] # needed for legacy reasons
+              [HashWrapper.new(result), hit]
+            end
+        end
+
+       [results, missing_records]
+      end
+    end
+
+    def build_hits
+      @build_hits ||= begin
+        if missing_records.any?
+          Searchkick.warn("Records in search index do not exist in database: #{missing_records.map { |v| "#{v[:model].model_name} #{v[:id]}" }.join(", ")}")
+        end
+        with_hit_and_missing_records[0]
+      end
+    end
+
     def results_query(records, hits)
+      records = Searchkick.scope(records)
+
       ids = hits.map { |hit| hit["_id"] }
       if options[:includes] || options[:model_includes]
         included_relations = []
         combine_includes(included_relations, options[:includes])
         combine_includes(included_relations, options[:model_includes][records]) if options[:model_includes]
 
-        records =
-          if defined?(NoBrainer::Document) && records < NoBrainer::Document
-            if Gem.loaded_specs["nobrainer"].version >= Gem::Version.new("0.21")
-              records.eager_load(included_relations)
-            else
-              records.preload(included_relations)
-            end
-          else
-            records.includes(included_relations)
-          end
+        records = records.includes(included_relations)
+      end
+
+      if options[:scope_results]
+        records = options[:scope_results].call(records)
       end
 
       Searchkick.load_records(records, ids)
@@ -231,5 +342,13 @@ module Searchkick
     def base_field(k)
       k.sub(/\.(analyzed|word_start|word_middle|word_end|text_start|text_middle|text_end|exact)\z/, "")
     end
+
+    def hit_highlights(hit, multiple: false)
+      if hit["highlight"]
+        hit["highlight"].to_h { |k, v| [(options[:json] ? k : k.sub(/\.#{@options[:match_suffix]}\z/, "")).to_sym, multiple ? v : v.first] }
+      else
+        {}
+      end
+    end
   end
 end

data/lib/searchkick/version.rb

@@ -1,3 +1,3 @@
 module Searchkick
-  VERSION = "2.3.2"
+  VERSION = "5.2.1"
 end

data/lib/searchkick/where.rb

@@ -0,0 +1,11 @@
+module Searchkick
+  class Where
+    def initialize(relation)
+      @relation = relation
+    end
+
+    def not(value)
+      @relation.where(_not: value)
+    end
+  end
+end