Merge pull request #169 from MITLibraries/tco-72

Adds barcode detector and lookup

Author: matt-bernhardt

app/graphql/types/standard_identifiers_type.rb

@@ -2,16 +2,18 @@
 
 module Types
   class StandardIdentifiersType < Types::BaseObject
-    description 'A detector for standard identifiers in search terms. Currently supported: ISBN, ISSN, PMID, DOI'
+    description 'A detector for standard identifiers in search terms. Currently supported: Barcode, ISBN, ISSN, PMID, DOI'
 
     field :details, DetailsType, description: 'Additional information about the detected identifier(s)'
-    field :kind, String, null: false, description: 'The type of identifier detected (one of ISBN, ISSN, PMID, DOI)'
+    field :kind, String, null: false, description: 'The type of identifier detected (one of Barcode, ISBN, ISSN, PMID, DOI)'
     field :value, String, null: false, description: 'The identifier detected in the search term'
 
     # details does external lookups and should only be run if the fields
     # have been explicitly requested
     def details
       case @object[:kind]
+      when :barcode
+        LookupBarcode.new.info(@object[:value])
       when :doi
         LookupDoi.new.info(@object[:value])
       when :isbn

app/models/detector/standard_identifiers.rb

@@ -1,7 +1,7 @@
 # frozen_string_literal: true
 
 class Detector
-  # Detector::StandardIdentifiers detects the identifiers DOI, ISBN, ISSN, PMID.
+  # Detector::StandardIdentifiers detects the identifiers Barcode, DOI, ISBN, ISSN, PMID.
   # See /docs/reference/pattern_detection_and_enhancement.md for details.
   class StandardIdentifiers
     attr_reader :detections
@@ -52,6 +52,7 @@ def self.record(term)
     # patterns are regex patterns to be applied to the basic search box input
     def patterns
       {
+        barcode: /^39080[0-9]{9}$/,
         isbn: /\b(ISBN-*(1[03])* *(: ){0,1})*(([0-9Xx][- ]*){13}|([0-9Xx][- ]*){10})\b/,
         issn: /\b[0-9]{4}-[0-9]{3}[0-9xX]\b/,
         pmid: /\b((pmid|PMID):\s?(\d{7,8}))\b/,

app/models/lookup_barcode.rb

@@ -0,0 +1,66 @@
+# frozen_string_literal: true
+
+# LookupBarcode takes a 14-digit integer (flagged by a regex within the Detector::StandardIdentifier class) and consults
+# the Primo API for the associated record. The structure of this class is pretty close to the other lookup models, with
+# an info method being the only public method. If Primo finds a record for the submitted barcode, the class returns some
+# metadata about the record, along with a link to the complete record using the discovery/fulldisplay path.
+class LookupBarcode
+  # info takes a barcode as an argument and returns associated metadata about that item, provided Primo is able to
+  # locate it. If no record is found for any reason, the method returns nil.
+  #
+  # @note While the barcode argument is technically a string, in reality it should be a 14-digit integer in order to
+  #       return anything meaningful.
+  # @param barcode String
+  # @return Hash or Nil
+  def info(barcode)
+    xml = fetch(barcode)
+
+    return if xml == 'Error'
+
+    metadata = extract_metadata(xml)
+
+    if metadata.reject { |_k, v| v.empty? }.present?
+      metadata[:barcode] = barcode
+      metadata[:link_resolver_url] = link_resolver_url(metadata)
+      metadata
+    else
+      Rails.logger.debug { "Barcode lookup error. Barcode #{barcode} detected by Primo returned no data" }
+      nil
+    end
+  end
+
+  private
+
+  def extract_metadata(xml)
+    {
+      recordId: xml.xpath('//default:recordIdentifier', 'default' => 'http://www.loc.gov/zing/srw/').text,
+      title: xml.xpath('//dc:title', 'dc' => 'http://purl.org/dc/elements/1.1/').text,
+      date: xml.xpath('//dc:date', 'dc' => 'http://purl.org/dc/elements/1.1/').text,
+      publisher: xml.xpath('//dc:publisher', 'dc' => 'http://purl.org/dc/elements/1.1/').text,
+      authors: xml.xpath('//dc:contributor', 'dc' => 'http://purl.org/dc/elements/1.1/').text
+    }
+  end
+
+  def url(barcode)
+    "https://mit.alma.exlibrisgroup.com/view/sru/01MIT_INST?version=1.2&operation=searchRetrieve&recordSchema=dc&query=alma.all_for_ui=#{barcode}"
+  end
+
+  def fetch(barcode)
+    resp = HTTP.headers(accept: 'application/xml').get(url(barcode))
+
+    if resp.status == 200
+      Nokogiri::XML(resp.to_s)
+    else
+      Rails.logger.debug do
+        "Barcode lookup error. Barcode #{barcode} detected but Primo returned an error status"
+      end
+      Rails.logger.debug { "URL: #{url(barcode)}" }
+      Sentry.capture_message('Primo API error after barcode detection')
+      'Error'
+    end
+  end
+
+  def link_resolver_url(metadata)
+    "https://mit.primo.exlibrisgroup.com/discovery/fulldisplay?vid=01MIT_INST:MIT&docid=alma#{metadata[:recordId]}"
+  end
+end

app/models/metrics/algorithms.rb

@@ -17,6 +17,7 @@
 #  suggested_resource_exact :integer
 #  lcsh                     :integer
 #  citation                 :integer
+#  barcode                  :integer
 #
 module Metrics
   # Algorithms aggregates statistics for matches for all SearchEvents
@@ -49,8 +50,8 @@ def generate(month = nil)
                 else
                   count_matches(SearchEvent.includes(:term))
                 end
-      Metrics::Algorithms.create(month:, citation: matches[:citation], doi: matches[:doi], issn: matches[:issn],
-                                 isbn: matches[:isbn], lcsh: matches[:lcsh], pmid: matches[:pmid],
+      Metrics::Algorithms.create(month:, barcode: matches[:barcode], citation: matches[:citation], doi: matches[:doi],
+                                 issn: matches[:issn], isbn: matches[:isbn], lcsh: matches[:lcsh], pmid: matches[:pmid],
                                  journal_exact: matches[:journal_exact],
                                  suggested_resource_exact: matches[:suggested_resource_exact],
                                  unmatched: matches[:unmatched])
@@ -122,7 +123,7 @@ def match_lcsh(event, matches)
     # @param matches [Hash] a Hash that keeps track of how many of each algorithm we match
     # @return [Array] an array of matched StandardIdentifiers
     def match_standard_identifiers(event, matches)
-      known_ids = %i[unmatched pmid isbn issn doi]
+      known_ids = %i[unmatched pmid isbn issn doi barcode]
       ids = Detector::StandardIdentifiers.new(event.term.phrase)
 
       known_ids.each do |id|

app/views/report/algorithm_metrics.html.erb

@@ -15,6 +15,7 @@
     <% else %>
       <th>Month</th>
     <% end %>
+    <th>Barcode</th>
     <th>DOI</th>
     <th>ISSN</th>
     <th>ISBN</th>
@@ -31,6 +32,7 @@
     <% else %>
       <td><%= metric.month.strftime("%B %Y") %></td>
     <% end %>
+    <td><%= metric.barcode %></td>
     <td><%= metric.doi %></td>
     <td><%= metric.issn %></td>
     <td><%= metric.isbn %></td>

db/migrate/20250110164046_add_barcode_to_metrics_algorithms.rb

@@ -0,0 +1,5 @@
+class AddBarcodeToMetricsAlgorithms < ActiveRecord::Migration[7.1]
+  def change
+    add_column :metrics_algorithms, :barcode, :integer
+  end
+end

db/schema.rb

@@ -41,8 +41,14 @@
 Detector.find_or_create_by(name: 'Journal')
 Detector.find_or_create_by(name: 'SuggestedResource')
 Detector.find_or_create_by(name: 'Citation')
+Detector.find_or_create_by(name: 'Barcode')
 
 # DetectorCategories
+DetectorCategory.find_or_create_by(
+  detector: Detector.find_by(name: 'Barcode'),
+  category: Category.find_by(name: 'Transactional'),
+  confidence: 0.95
+)
 DetectorCategory.find_or_create_by(
   detector: Detector.find_by(name: 'Citation'),
   category: Category.find_by(name: 'Transactional'),

db/seeds.rb

@@ -72,6 +72,30 @@ class GraphqlControllerTest < ActionDispatch::IntegrationTest
     assert_equal('10.1038/nphys1170', json['data']['logSearchEvent']['detectors']['standardIdentifiers'].first['value'])
   end
 
+  test 'search event query can return detected barcodes' do
+    VCR.use_cassette('barcode 39080027236626') do
+      post '/graphql', params: { query: '{
+        logSearchEvent(sourceSystem: "timdex", searchTerm: "39080027236626") {
+          detectors {
+            standardIdentifiers {
+              kind
+              value
+              details {
+                title
+              }
+            }
+          }
+        }
+      }' }
+
+      json = response.parsed_body
+
+      assert_equal('barcode', json['data']['logSearchEvent']['detectors']['standardIdentifiers'].first['kind'])
+      assert_equal('39080027236626', json['data']['logSearchEvent']['detectors']['standardIdentifiers'].first['value'])
+      assert_equal('Transactions of the Institution of Naval Architects.', json['data']['logSearchEvent']['detectors']['standardIdentifiers'].first['details']['title'])
+    end
+  end
+
   test 'search event query can return detected journals' do
     post '/graphql', params: { query: '{
                                  logSearchEvent(sourceSystem: "timdex", searchTerm: "nature") {

test/controllers/graphql_controller_test.rb

@@ -43,3 +43,8 @@ seven:
   detector: citation
   category: transactional
   confidence: 0.3
+
+eight:
+  detector: barcode
+  category: transactional
+  confidence: 0.95

test/fixtures/detector_categories.yml

@@ -7,6 +7,9 @@
 #  created_at :datetime         not null
 #  updated_at :datetime         not null
 #
+barcode:
+  name: 'Barcode'
+
 citation:
   name: 'Citation'
 

test/fixtures/detectors.yml

@@ -39,3 +39,9 @@ multiple_detections:
 
 citation:
   value: '12 2 2005 2007 6 a accessed altun available context current dec education experience httpcieedasueduvolume6number12 hypertext in issues july language learners no of on online reading serial the understanding vol web'
+
+barcode:
+  value: 39080678901234
+
+not_a_barcode:
+  value: '39080678901234 extra some text with'

test/fixtures/fingerprints.yml

@@ -55,3 +55,10 @@ old_suggested_resource_jstor:
   term: suggested_resource_jstor
   source: test
   created_at: <%= 1.year.ago %>
+current_month_barcode:
+  term: barcode
+  source: test
+old_barcode:
+  term: barcode
+  source: test
+  created_at: <%= 1.year.ago %>

test/fixtures/search_events.yml

@@ -57,3 +57,11 @@ multiple_detections:
 citation:
   phrase: "A. Altun, "Understanding hypertext in the context of reading on the web: Language learners' experience," Current Issues in Education, vol. 6, no. 12, July, 2005. [Online serial]. Available: http://cie.ed.asu.edu/volume6/number12/. [Accessed Dec. 2, 2007]."
   fingerprint: citation
+
+barcode:
+  phrase: '39080678901234'
+  fingerprint: barcode
+
+not_a_barcode:
+  phrase: '39080678901234 with some extra text'
+  fingerprint: not_a_barcode

test/fixtures/terms.yml

@@ -25,7 +25,7 @@ class CitationTest < ActiveSupport::TestCase
     test 'standard_identifier_bulk_checker' do
       bulk = Detector::StandardIdentifiers.check_all_matches(output: true)
 
-      assert_equal(5, bulk.count)
+      assert_equal(6, bulk.count)
     end
 
     test 'suggested_resources_bulk_checker' do

test/models/detector/bulk_checker_test.rb

@@ -0,0 +1,35 @@
+# frozen_string_literal: true
+
+require 'test_helper'
+
+class LookupBarcodeTest < ActiveSupport::TestCase
+  test 'metadata object is returned with expected fields' do
+    VCR.use_cassette('barcode 39080027236626') do
+      metadata = LookupBarcode.new.info('39080027236626')
+
+      expected_keys = %i[title date publisher authors link_resolver_url]
+
+      expected_keys.each do |key|
+        assert_includes(metadata.keys, key)
+      end
+    end
+  end
+
+  test 'link resolver URL returns a simple item URL' do
+    VCR.use_cassette('barcode 39080027236626') do
+      metadata = LookupBarcode.new.info('39080027236626')
+
+      expected_url = 'https://mit.primo.exlibrisgroup.com/discovery/fulldisplay?vid=01MIT_INST:MIT&docid=alma990002933430106761'
+
+      assert_equal(expected_url, metadata[:link_resolver_url])
+    end
+  end
+
+  test 'barcode not found' do
+    VCR.use_cassette('barcode not found') do
+      metadata = LookupBarcode.new.info('this-is-not-a-barcode')
+
+      assert_nil(metadata)
+    end
+  end
+end

test/models/lookup_barcode_test.rb

@@ -17,11 +17,18 @@
 #  suggested_resource_exact :integer
 #  lcsh                     :integer
 #  citation                 :integer
+#  barcode                  :integer
 #
 require 'test_helper'
 
 class Algorithms < ActiveSupport::TestCase
   # Monthlies
+  test 'barcode counts are included in monthly aggregation' do
+    aggregate = Metrics::Algorithms.new.generate(DateTime.now)
+
+    assert_equal 1, aggregate.barcode
+  end
+
   test 'citation counts are included in monthly aggregation' do
     aggregate = Metrics::Algorithms.new.generate(DateTime.now)
 
@@ -92,6 +99,11 @@ class Algorithms < ActiveSupport::TestCase
     # drop all searchevents to make math easier and minimize fragility over time as more fixtures are created
     SearchEvent.delete_all
 
+    barcode_expected_count = rand(1...100)
+    barcode_expected_count.times do
+      SearchEvent.create(term: terms(:barcode), source: 'test')
+    end
+
     citation_expected_count = rand(1...100)
     citation_expected_count.times do
       SearchEvent.create(term: terms(:citation), source: 'test')
@@ -129,6 +141,7 @@ class Algorithms < ActiveSupport::TestCase
 
     aggregate = Metrics::Algorithms.new.generate(DateTime.now)
 
+    assert_equal barcode_expected_count, aggregate.barcode
     assert_equal citation_expected_count, aggregate.citation
     assert_equal doi_expected_count, aggregate.doi
     assert_equal issn_expected_count, aggregate.issn
@@ -139,6 +152,12 @@ class Algorithms < ActiveSupport::TestCase
   end
 
   # Total
+  test 'barcode counts are included in total aggregation' do
+    aggregate = Metrics::Algorithms.new.generate
+
+    assert_equal 2, aggregate.barcode
+  end
+
   test 'citation counts are included in total aggregation' do
     aggregate = Metrics::Algorithms.new.generate
 
@@ -197,6 +216,11 @@ class Algorithms < ActiveSupport::TestCase
     # drop all searchevents to make math easier and minimize fragility over time as more fixtures are created
     SearchEvent.delete_all
 
+    barcode_expected_count = rand(1...100)
+    barcode_expected_count.times do
+      SearchEvent.create(term: terms(:barcode), source: 'test')
+    end
+
     citation_expected_count = rand(1...100)
     citation_expected_count.times do
       SearchEvent.create(term: terms(:citation), source: 'test')
@@ -239,6 +263,7 @@ class Algorithms < ActiveSupport::TestCase
 
     aggregate = Metrics::Algorithms.new.generate
 
+    assert_equal barcode_expected_count, aggregate.barcode
     assert_equal citation_expected_count, aggregate.citation
     assert_equal doi_expected_count, aggregate.doi
     assert_equal issn_expected_count, aggregate.issn

test/models/metrics/algorithms_test.rb

@@ -38,6 +38,11 @@
     header&.each do |redacted_text|
       interaction.filter!(redacted_text, '<REDACTED_NEL>')
     end
+
+    header = interaction.response&.headers&.[]('Set-Cookie')
+    header&.each do |redacted_text|
+      interaction.filter!(redacted_text, '<FAKE_COOKIE_DATA>')
+    end
   end
 end