Replace our custom code with new traject delete_if.

* Not having to maintain custom code is a good thing.
tulibraries · Feb 9, 2022 · 02d1cea · 02d1cea
1 parent 4627685
commit 02d1cea
Show file tree

Hide file tree

Showing 6 changed files with 31 additions and 101 deletions.
diff --git a/Gemfile.lock b/Gemfile.lock
@@ -166,7 +166,7 @@ GEM
       guard (~> 2.1)
       guard-compat (~> 1.1)
       rspec (>= 2.99.0, < 4.0)
-    hashie (4.1.0)
+    hashie (5.0.0)
     http (5.0.4)
       addressable (~> 2.8)
       http-cookie (~> 1.0)
@@ -331,10 +331,10 @@ GEM
       activesupport (>= 5.2)
       sprockets (>= 3.0.0)
     thor (1.1.0)
-    traject (3.6.0)
+    traject (3.7.0)
       concurrent-ruby (>= 0.8.0)
       dot-properties (>= 0.1.1)
-      hashie (>= 3.1, < 5)
+      hashie (>= 3.1, < 6)
       http (>= 3.0, < 6)
       httpclient (~> 2.5)
       marc (~> 1.0)

diff --git a/lib/cob_index/indexer_config.rb b/lib/cob_index/indexer_config.rb
@@ -16,9 +16,6 @@
 # Include custom traject macros
 extend CobIndex::Macros::Custom
 
-# include unicode normalize for thread safety
-extend CobIndex::Macros::Transformations
-
 CORPORATE_NAMES = CobIndex::Util.load_list_file("corporate_names")
 
 settings(&CobIndex::DefaultConfig.indexer_settings)
@@ -80,7 +77,7 @@
 
 # Creator/contributor fields
 to_field "creator_t", extract_marc_with_flank("245c:100abcdejlmnopqrtu:110abcdelmnopt:111acdejlnopt:700abcdejqu:710abcde:711acdej", trim_punctuation: true)
-to_field "creator_facet", extract_marc("100abcdq:110abcd:111ancdj:700abcdq:710abcd:711ancdj", trim_punctuation: true), filter_values(CORPORATE_NAMES)
+to_field "creator_facet", extract_marc("100abcdq:110abcd:111ancdj:700abcdq:710abcd:711ancdj", trim_punctuation: true), delete_if(CORPORATE_NAMES)
 to_field "creator_display", extract_creator
 to_field "contributor_display", extract_contributor
 to_field "creator_vern_display", extract_creator_vern
@@ -201,6 +198,7 @@
 to_field "items_json_display", extract_item_info
 
 # Identifier fields
+# include unicode normalize for thread safety
 to_field("isbn_display",  extract_marc("020a", separator: nil), &normalize_isbn)
 to_field("alt_isbn_display",  extract_marc("020z:776z", separator: nil), &normalize_isbn)
 to_field("issn_display", extract_marc("022a", separator: nil), &normalize_issn)

diff --git a/lib/cob_index/macros.rb b/lib/cob_index/macros.rb
@@ -1,7 +1,6 @@
 # frozen_string_literal: true
 
 module CobIndex::Macros
-  autoload :Transformations, "cob_index/macros/transformations"
   autoload :Custom, "cob_index/macros/custom"
   autoload :Marc21, "cob_index/macros/marc21"
   autoload :MarcFormats, "cob_index/macros/marc_format_classifier"

diff --git a/lib/cob_index/macros/transformations.rb b/lib/cob_index/macros/transformations.rb
diff --git a/spec/cob_index/indexer_config_spec.rb b/spec/cob_index/indexer_config_spec.rb
@@ -140,4 +140,30 @@
       end
     end
   end
+
+  describe "creator_facet field" do
+    before do
+      stub_const("ENV", ENV.to_hash.merge("SOLR_URL" => "foo"))
+      indexer.load_config_file("lib/cob_index/indexer_config.rb")
+    end
+
+    context "Unwanted corp present" do
+      # Note: that "Books24x7, Inc" is excluded via ./lib/list/corporate_names.txt
+      let(:record_text) { "
+        <record>
+          <datafield ind1=' ' ind2=' ' tag='100'>
+            <subfield code='b'>FOO</subfield>
+          </datafield>
+          <datafield ind1=' ' ind2=' ' tag='100'>
+            <subfield code='a'>Books24x7, Inc</subfield>
+          </datafield>
+        </record>
+      " }
+
+      it "it removes the unwanted corp" do
+        expect(indexer.map_record(record)["creator_facet"]).to eq(["FOO"])
+      end
+    end
+
+  end
 end
diff --git a/spec/cob_index/macros/transformations_spec.rb b/spec/cob_index/macros/transformations_spec.rb