HYC-1951 - Refine Dimensions Query (#1116)

* prioritize unc author affiliations

* update query string to catch all variances of UNC and eliminate false positives

* author to hash tests

* slightly changed wording

* update test expectations

* update test expectations

app/services/tasks/dimensions_ingest_service.rb

@@ -4,6 +4,7 @@ module Tasks
   class DimensionsIngestService
     include Tasks::IngestHelper
     attr_reader :admin_set, :depositor
+    UNC_GRID_ID = 'grid.410711.2'
 
     def initialize(config)
       @config = config
@@ -104,11 +105,22 @@ def author_to_hash(author, index)
       }
       # Add first author affiliation to other affiliation array
       if author['affiliations'].present?
-        hash['other_affiliation'] = author['affiliations'][0]['raw_affiliation']
+        hash['other_affiliation'] = retrieve_author_affiliation(author['affiliations'])
       end
       hash
     end
 
+    def retrieve_author_affiliation(affiliations)
+      unc_affiliations = affiliations.select { |affiliation| affiliation['id'] == UNC_GRID_ID }
+      if !unc_affiliations.empty?
+        # Prioritize UNC affiliations, only retrieving the first one
+        return unc_affiliations[0]['raw_affiliation']
+      end
+      # Otherwise, retrieve the first affiliation
+      return affiliations[0]['raw_affiliation']
+    end
+
+
     def format_publication_identifiers(publication)
       [
         publication['id'].present? ? "Dimensions ID: #{publication['id']}" : nil,

app/services/tasks/dimensions_query_service.rb

@@ -197,10 +197,11 @@ def solr_query_builder(pub)
     def generate_query_string(start_date, end_date, page_size, cursor)
       search_clauses = ['where type = "article"', "date >= \"#{start_date}\"", "date < \"#{end_date}\""].join(' and ')
       return_fields = ['basics', 'extras', 'abstract', 'issn', 'publisher', 'journal_title_raw', 'linkout', 'concepts'].join(' + ')
+      unc_affiliation_variants = ['"UNC-CH"', '"University of North Carolina at Chapel Hill"', '"UNC-Chapel Hill"', '"University of North Carolina-Chapel Hill"', '"University of North Carolina, Chapel Hill"'].join(' OR ')
       <<~QUERY
         search publications #{search_clauses} in raw_affiliations
         for """
-        "University of North Carolina, Chapel Hill" OR "UNC"
+          #{unc_affiliation_variants}
         """
         return publications[#{return_fields}]
         limit #{page_size}

spec/fixtures/files/dimensions_ingest_test_fixture.json

@@ -98,7 +98,9 @@
                     "current_organization_id": "grid.10698.36",
                     "first_name": "Susan L",
                     "last_name": "Hogan",
-                    "orcid": null,
+                    "orcid": [
+                        "0000-0000-0000-0000"
+                    ],
                     "raw_affiliation": [
                         "UNC Kidney Center, Division of Nephrology and Hypertension, University of North Carolina, Chapel Hill."
                     ],

spec/services/tasks/dimensions_ingest_service_spec.rb

@@ -305,4 +305,59 @@
       expect(article.keyword).to eq([])
     end
   end
+
+  describe '#author_to_hash' do
+    let (:unc_grid_id)  { 'grid.410711.2' }
+    let (:non_unc_affiliation) {
+      {
+        'city' => 'Test City',
+        'city_id' => 5318313,
+        'country' => 'United States',
+        'country_code' =>  'US',
+        'id' =>  'grid.134563.6',
+        'name' =>  'Test University',
+        'raw_affiliation' =>  'Test Raw Affiliation',
+        'state' => 'Test-State',
+        'state_code' =>  'US-AZ'
+     }
+    }
+
+    context 'when an author has multiple affiliations' do
+      it 'uses their first affiliation to populate the author hash if no UNC affiliation exists' do
+        non_unc_affiliated_author = test_publications.first['authors'].find { |author| author['id'] != unc_grid_id }
+        # Ensure the author has multiple non-unc affiliations
+        non_unc_affiliated_author['affiliations'].append(non_unc_affiliation) unless non_unc_affiliated_author['affiliations'].size > 1
+        author_hash = service.author_to_hash(non_unc_affiliated_author, 0)
+        # Check that the author hash contains the expected metadata from the first affiliation
+        expect(author_hash).to eq(
+          {
+            'name' => 'Thorpe, Carolyn T',
+            'other_affiliation' => 'Eshelman School of Pharmacy, University of North Carolina, Chapel Hill.',
+            'orcid' => 'https://orcid.org/0000-0002-7662-7497',
+            'index' => '1'
+          }
+        )
+      end
+
+      it 'prioritizes retrieval of the UNC affiliation even if it is not the first one' do
+        first_publication_authors = test_publications.first['authors']
+        # Retrieve the first UNC-affiliated author and their first UNC-affiliation
+        unc_affiliated_author = first_publication_authors.find do |author|
+          author['affiliations'].any? { |affiliation| affiliation['id'] == unc_grid_id }
+        end
+        first_unc_affiliation = unc_affiliated_author['affiliations'].find { |affiliation| affiliation['id'] == unc_grid_id }
+        # Ensure the author's first affiliation is not the UNC affiliation
+        unc_affiliated_author['affiliations'].unshift(non_unc_affiliation)
+        author_hash = service.author_to_hash(unc_affiliated_author, 0)
+        expect(author_hash).to eq(
+          {
+            'name' => 'Hogan, Susan L',
+            'other_affiliation' => 'UNC Kidney Center, Division of Nephrology and Hypertension, University of North Carolina, Chapel Hill.',
+            'orcid' => 'https://orcid.org/0000-0000-0000-0000',
+            'index' => '1'
+          }
+        )
+      end
+    end
+  end
 end