uc3-dmp-id 0.1.37 → 0.1.38

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 4b8c0ff9aa6add1fc2da937d9f76872ae6bb287adf5df008d9c0d434f25fd717
-  data.tar.gz: a0479d353eb1d64cfab1bdb718f3efcf6a1d66cec8ea444ab735e74708f88984
+  metadata.gz: cec9b27f000aa3543c1a53c70c2ac6e6734ad1e0c70a98b7f583db986941ba43
+  data.tar.gz: da94a81c40adc7d9c4bea9767ec90bbdeab18f4445a27853fe1ae2e5d51b3c97
 SHA512:
-  metadata.gz: a578c6b89bb5d94172d5e2c24255f81326871616767039a4aa4de6d1c00da4a9d16fcf20f26a4479d8d1d5017e90dfcb27ca9c6c4bb5b1ed5421928bfb518eeb
-  data.tar.gz: 3dc8d1999fb6f26cdf8fc465d160513982ac7607bfb7c5e4c5f719b4f43d3b5d82261c5c4ce87cdd10be1d1631276b212b54cfd60b3f5523b389b3f4da25dbe5
+  metadata.gz: e89112bbaae3d06adb736c4b0d846c82b27dbfdf90db8d73079a5b83740f04b1f771f8214121884275fd907a8150ad640e9b85689320d23e628d466c8f0d0a65
+  data.tar.gz: a2a27c70445f104015bbe15cd5e19e2fffb0de82eae908ee1d335438d18cc29809ad334bdcb92280066541a22a5cb040b4ce1f3ab1d54cd0d9a217b8d6aeff92

data/lib/uc3-dmp-id/comparator.rb

@@ -70,14 +70,15 @@ module Uc3DmpId
         # Compare the people involved, their affiliations and any funding opportunity numbers
         response = _opportunities_match?(array: hash.fetch('funder_opportunity_ids', []), dmp:, response:)
         response = _orcids_match?(array: hash.fetch('people_ids', []), dmp:, response:)
-        response = _last_name_and_affiliation_match?(hash:, dmp:, response:)
+        response = _last_name_match?(hash:, dmp:, response:)
+        response = _affiliation_match?(hash:, dmp:, response:)
 
         # Only process the following if we had some matching people, affiliations or opportunity nbrs
         response = _repository_match?(hash:, dmp:, response:) if response[:score].positive?
         response = _text_match?(type: 'title', text: hash['title'], dmp:, response:) if response[:score].positive?
         response = _text_match?(type: 'abstract', text: hash['description'], dmp:, response:) if response[:score].positive?
-        # If the score is less than 4 then we have no confidence that it is a match
-        next if response[:score] <= 3
+        # If the score is less than 3 then we have no confidence that it is a match
+        next if response[:score] <= 2
 
         # Set the confidence level based on the score
         response[:dmp_id] = "DMP##{dmp['dmp_id']}"
@@ -178,29 +179,46 @@ module Uc3DmpId
     #     affiliation_ids: ["https://ror.org/blah"]
     #   }
     # rubocop:disable Metrics/AbcSize, Metrics/CyclomaticComplexity, Metrics/PerceivedComplexity
-    def _last_name_and_affiliation_match?(hash:, dmp:, response:)
+    def _last_name_match?(hash:, dmp:, response:)
       return response unless hash.is_a?(Hash) && dmp.is_a?(Hash) && response.is_a?(Hash)
       return response unless hash['people'].is_a?(Array)
 
       # Check the person last names and affiliation name and RORs
       last_names_matched = _compare_arrays(array_a: dmp['people'], array_b: hash['people'])
+      return response if last_names_matched <= 0
+
+      @logger&.debug(
+        message: 'Contributor name match',
+        details: {
+          people: { dmp: dmp['people'], work: hash['people'] }
+        }
+      )
+      response[:score] += last_names_matched * 2
+      response[:notes] << 'contributor names matched'
+      response
+    end
+    # rubocop:enable Metrics/AbcSize, Metrics/CyclomaticComplexity, Metrics/PerceivedComplexity
+
+    def _affiliation_match?(hash:, dmp:, response:)
+      return response unless hash.is_a?(Hash) && dmp.is_a?(Hash) && response.is_a?(Hash)
+      return response unless hash['affiliations'].is_a?(Array) || hash['affiliation_ids'].is_a?(Array)
+
+      # Check the affiliation names and RORs
       rors_matched = _compare_arrays(array_a: dmp.fetch('affiliation_ids', []), array_b: hash['affiliation_ids'])
       affil_names_matched = _compare_arrays(array_a: dmp.fetch('affiliations', []), array_b: hash['affiliations'])
-      return response if last_names_matched <= 0 && rors_matched <= 0 && affil_names_matched <= 0
+      return response if rors_matched <= 0 && affil_names_matched <= 0
 
       @logger&.debug(
-        message: 'Contributor name match',
+        message: 'Affiliation name match',
         details: {
-          people: { dmp: dmp['people'], work: hash['people'] },
           rors: { dmp: dmp['affiliation_ids'], work: hash['affiliation_ids'] },
           places: { dmp: dmp['affiliations'], work: hash['affiliations'] }
         }
       )
-      response[:score] += last_names_matched + rors_matched + affil_names_matched
-      response[:notes] << 'contributor names and/or affiliations matched'
+      response[:score] += rors_matched + affil_names_matched
+      response[:notes] << 'affiliations matched'
       response
     end
-    # rubocop:enable Metrics/AbcSize, Metrics/CyclomaticComplexity, Metrics/PerceivedComplexity
 
     # Returns whether or not the incoming list of repositories match those defined in the DMP. Expecting:
     #    {

data/lib/uc3-dmp-id/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Uc3DmpId
-  VERSION = '0.1.37'
+  VERSION = '0.1.38'
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: uc3-dmp-id
 version: !ruby/object:Gem::Version
-  version: 0.1.37
+  version: 0.1.38
 platform: ruby
 authors:
 - Brian Riley
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2024-05-02 00:00:00.000000000 Z
+date: 2024-05-06 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: json