Implemented the EngineeringSearch experimental SearchBuilder for

supporting a different Solr RequestHandler

Co-authored-by: Anna Headley <[email protected]>
Co-authored-by: Carolyn Cole <[email protected]>
Co-authored-by: Eliot Jordan <[email protected]>
Co-authored-by: James R. Griffin III <[email protected]>
Co-authored-by: Trey Pendragon <[email protected]>

app/controllers/catalog_controller.rb

@@ -787,4 +787,27 @@ def basic_response
       render plain: 'OK'
       nil
     end
+
+    def search_service_context
+      return {} unless Flipflop.multi_algorithm?
+      return {} unless alternate_search_builder_class # use default if none specified
+      { search_builder_class: alternate_search_builder_class }
+    end
+
+    def alternate_search_builder_class
+      return unless search_algorithm_param && allowed_search_algorithms.include?(search_algorithm_param)
+
+      "#{search_algorithm_param}_search_builder".camelize.constantize
+    end
+
+    # When adding a new Ranking Algorithm the name will need to be added to this list before it can be utilized
+    # For example if you added a cats ranking algorithm, with a CatsSearchBuilder you would add "cats" to this list ["engineering","cats"]
+    # This is to make sure the user can not just execute any SearchBuilder in the system
+    def allowed_search_algorithms
+      ["engineering"]
+    end
+
+    def search_algorithm_param
+      params[:search_algorithm]
+    end
 end

app/models/engineering_search_builder.rb

@@ -0,0 +1,10 @@
+# frozen_string_literal: true
+
+# This is a demonstration search builder, not intended for prodcution use.
+class EngineeringSearchBuilder < SearchBuilder
+  self.default_processor_chain += [:switch_request_handler]
+
+  def switch_request_handler(solr_parameters)
+    solr_parameters[:qt] = "engineering_search"
+  end
+end

solr/conf/solrconfig.xml

@@ -361,6 +361,242 @@
     </lst>
   </requestHandler>
 
+  <requestHandler name="engineering_search" class="solr.SearchHandler">
+    <!-- default values for query parameters can be specified, these
+         will be overridden by parameters in the request
+      -->
+    <lst name="defaults">
+      <str name="defType">edismax</str>
+      <str name="echoParams">explicit</str>
+      <int name="rows">10</int>
+      <str name="sow">on</str>
+      <str name="q.alt">*:*</str>
+      <str name="mm">6&lt;90%</str>
+
+      <!-- boost local holdings (add 50 to the score) to reduce unnecessary requests -->
+      <str name="bf">if(field(numeric_id_b),50,0)</str>
+      <str name="boost">if(termfreq(text,'engineering'),100,0)</str>
+
+       <!-- this qf and pf are used by default, if not otherwise specified by
+            client. The default blacklight_config will use these for the
+            "keywords" search. See the author_qf/author_pf, title_qf, etc
+            below, which the default blacklight_config will specify for
+            those searches. You may also be interested in:
+            http://wiki.apache.org/solr/LocalParams
+       -->
+
+      <str name="qf">
+        title_a_index^1500
+        author_main_unstem_search^1000
+        title_unstem_search^40
+        title_display^40
+        author_unstem_search^40
+        subject_topic_unstem_search^18
+        subject_unstem_search^15
+        siku_subject_unstem_search^15
+        local_subject_unstem_search^15
+        homoit_subject_unstem_search^15
+        subject_topic_index^12
+        genre_unstem_search^10
+        subject_t^10
+        subject_addl_unstem_search^8
+        subject_addl_t^4
+        isbn_t^3
+        issn_s^3
+        lccn_s^3
+        uncontrolled_keyword_unstem_search^3
+        text
+        description_t
+        cjk_all
+        cjk_text
+      </str>
+      <str name="pf">
+        title_245a_lr^16000
+        title_245_lr^16000
+        title_a_index^12000
+        author_main_unstem_search^10000
+        title_unstem_search^400
+        title_display^400
+        author_unstem_search^400
+        subject_topic_unstem_search^180
+        subject_unstem_search^150
+        siku_subject_unstem_search^150
+        local_subject_unstem_search^150
+        homoit_subject_unstem_search^150
+        subject_topic_index^120
+        genre_unstem_search^100
+        subject_t^100
+        subject_addl_unstem_search^80
+        subject_addl_t^40
+        isbn_t^30
+        issn_s^30
+        lccn_s^30
+        uncontrolled_keyword_unstem_search^10
+        text^10
+        description_t^10
+        cjk_all^10
+        cjk_text^10
+      </str>
+      <str name="author_qf">
+        author_main_unstem_search^20
+        author_unstem_search^10
+        cjk_author
+      </str>
+      <str name="author_pf">
+        author_main_unstem_search^200
+        author_unstem_search^100
+        cjk_author^10
+      </str>
+      <str name="left_anchor_qf">
+        title_245a_la^50
+        title_245_la^10
+        title_la^2
+        title_addl_la
+      </str>
+      <str name="left_anchor_pf">
+        title_245a_lr^600
+        title_245_lr^600
+        title_245a_la^500
+        title_245_la^100
+        title_lr^100
+        title_la^20
+        title_addl_la^10
+      </str>
+      <str name="in_series_qf">
+        more_in_this_series_la
+      </str>
+      <str name="in_series_pf">
+        more_in_this_series_la
+      </str>
+      <str name="publisher_qf">
+        pub_created_unstem_search
+        cjk_publisher
+      </str>
+      <str name="publisher_pf">
+        pub_created_unstem_search
+        cjk_publisher
+      </str>
+      <str name="notes_qf">
+        notes_index
+        cjk_notes
+        cjk_notes_copied
+      </str>
+      <str name="notes_pf">
+        notes_index
+        cjk_notes
+        cjk_notes_copied
+      </str>
+      <str name="series_title_qf">
+        series_title_index^5
+        series_ae_index
+        series_statement_index
+        linked_series_title_index
+        linked_series_index
+        original_version_series_index
+        cjk_series_title
+      </str>
+      <str name="series_title_pf">
+        series_title_index^50
+        series_ae_index^10
+        series_statement_index^10
+        linked_series_title_index^10
+        linked_series_index^10
+        original_version_series_index^10
+        cjk_series_title^10
+      </str>
+      <str name="title_qf">
+        title_a_index^500
+        title_unstem_search^100
+        title_display^50
+        other_title_index^5
+        series_title_index^5
+        uniform_title_s^5
+        title_vern_display
+        content_title_index
+        contains_title_index
+        linked_title_index
+        series_ae_index
+        series_statement_index
+        linked_series_title_index
+        linked_series_index
+        original_version_series_index
+        cjk_title
+      </str>
+      <str name="title_pf">
+        title_245a_lr^5500
+        title_245_lr^5500
+        title_a_index^5000
+        title_unstem_search^1000
+        title_display^500
+        other_title_index^50
+        series_title_index^50
+        uniform_title_s^50
+        title_vern_display^10
+        content_title_index^10
+        contains_title_index^10
+        linked_title_index^10
+        series_ae_index^10
+        series_statement_index^10
+        linked_series_title_index^10
+        linked_series_index^10
+        original_version_series_index^10
+        cjk_title^10
+      </str>
+      <str name="subject_qf">
+        subject_topic_unstem_search^25
+        subject_unstem_search^20
+        genre_unstem_search^15
+        siku_subject_unstem_search
+        local_subject_unstem_search
+        homoit_subject_unstem_search
+        cjk_subject
+      </str>
+      <str name="subject_pf">
+        subject_topic_unstem_search^250
+        subject_unstem_search^200
+        genre_unstem_search^150
+        siku_subject_unstem_search^10
+        local_subject_unstem_search^10
+        homoit_subject_unstem_search^10
+        cjk_subject^10
+      </str>
+
+      <int name="ps">3</int>
+      <float name="tie">0.01</float>
+
+      <str name="fl">
+        id,
+        score,
+        author_display,
+        marc_relator_display,
+        format,
+        pub_created_display,
+        title_display,
+        title_vern_display,
+        isbn_s,
+        oclc_s,
+        lccn_s,
+        holdings_1display,
+        electronic_access_1display,
+        electronic_portfolio_s,
+        cataloged_tdt,
+        contained_in_s
+      </str>
+
+      <str name="facet">true</str>
+      <str name="facet.mincount">1</str>
+      <str name="facet.limit">10</str>
+      <str name="facet.field">format</str>
+      <str name="facet.field">language_facet</str>
+      <str name="facet.field">pub_date_start_sort</str>
+      <str name="facet.field">advanced_location_s</str>
+      <str name="f.format.facet.sort">index</str>
+      <str name="f.advanced_location_s.facet.sort">index</str>
+      <str name="f.language_facet.facet.limit">1000</str>
+      <str name="f.advanced_location_s.facet.limit">500</str>
+    </lst>
+  </requestHandler>
+
   <requestHandler name="/advanced" class="solr.SearchHandler">
     <!-- a lucene request handler for using the JSON Query DSL, for advanced search
          Using a separate request handler as a work around to

spec/controllers/catalog_controller_spec.rb

@@ -117,4 +117,75 @@
       end
     end
   end
+
+  describe "switching search algorithms" do
+    let(:repository) { CatalogController.blacklight_config.repository }
+
+    before do
+      allow(repository).to receive(:search).and_call_original
+      allow(controller.blacklight_config).to receive(:repository).and_return(repository)
+    end
+
+    context "when the multi-algorithm feature is on" do
+      before do
+        allow(Flipflop).to receive(:multi_algorithm?).and_return(true)
+      end
+
+      context "when the search_algorithm parameter is not present" do
+        it "uses the default search builder" do
+          get :index, params: { q: "coffee" }
+
+          expect(repository).to have_received(:search).with(instance_of(SearchBuilder))
+        end
+      end
+
+      context "when the search_algorithm parameter is set to 'engineering'" do
+        it "uses the alternate search builder" do
+          get :index, params: { q: "coffee", search_algorithm: "engineering" }
+
+          expect(repository).to have_received(:search).with(instance_of(EngineeringSearchBuilder))
+        end
+      end
+
+      context "when the search_algorithm parameter is set to 'not_a_real_class'" do
+        it "uses the default search builder" do
+          get :index, params: { q: "coffee", search_algorithm: "not_a_real_class" }
+
+          expect(repository).to have_received(:search).with(instance_of(SearchBuilder))
+        end
+      end
+
+      context "when the default is not search builder and the search_algorithm parameter is empty" do
+        it "uses the configured search builder" do
+          allow(controller.blacklight_config).to receive(:search_builder_class).and_return(EngineeringSearchBuilder)
+
+          get :index, params: { q: "coffee" }
+
+          expect(repository).to have_received(:search).with(instance_of(EngineeringSearchBuilder))
+        end
+      end
+    end
+
+    context "when the multi-algorithm feature is off" do
+      before do
+        allow(Flipflop).to receive(:multi_algorithm?).and_return(false)
+      end
+
+      context "when the search_algorithm parameter is not present" do
+        it "uses the default search builder" do
+          get :index, params: { q: "coffee" }
+
+          expect(repository).to have_received(:search).with(instance_of(SearchBuilder))
+        end
+      end
+
+      context "when the search_algorithm parameter is set to engineering" do
+        it "uses the alternate search builder" do
+          get :index, params: { q: "coffee", search_algorithm: "engineering" }
+
+          expect(repository).to have_received(:search).with(instance_of(SearchBuilder))
+        end
+      end
+    end
+  end
 end

spec/requests/request_spec.rb

@@ -543,4 +543,28 @@
       expect(json["data"][0]["attributes"]["electronic_portfolio_s"]).not_to be_blank
     end
   end
+
+  describe 'search algorithm selection' do
+    before do
+      allow(Flipflop).to receive(:multi_algorithm?).and_return(true)
+    end
+
+    context "when the search_algorithm parameter is not present" do
+      it "ranks using the default request handler" do
+        get "/catalog.json?q=roman"
+        json = JSON.parse(response.body)
+
+        expect(json["data"][0]["attributes"]["title"]).to eq "Ogonek : roman / ."
+      end
+    end
+
+    context "when the search_algorithm parameter is set to 'engineering'" do
+      it "ranks using the engineering request handler" do
+        get "/catalog.json?q=roman&search_algorithm=engineering"
+        json = JSON.parse(response.body)
+
+        expect(json["data"][0]["attributes"]["title"]).to eq "Reconstructing the Vitruvian Scorpio: An Engineering Analysis of Roman Field Artillery"
+      end
+    end
+  end
 end