uc3-dmp-id 0.1.50 → 0.1.52

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 7e3c4a8acb59e69dd223f10dab9ce2bcf8e7fa2ad131eed5bd6f04feced6e668
-  data.tar.gz: f8ad221d21001a798415ecced5dff904d20c5ca2d1e7021a7ed28a49f496f57f
+  metadata.gz: a126b7c8bbcf6836334748edef7a8ae44510b1c980c208de7448fffbb50f39df
+  data.tar.gz: d597f5f334d488e9dd1d1c25d475949daf9eb01b2a44d0166bdb8cb6f106bc4c
 SHA512:
-  metadata.gz: 84157070699350225aac0549b2e5a2879dc3ee95e33aa7e9139124f071109e8a51ea6c7789c68520a52f498dfbf5cee24475cddc5c0b6724decbee60feb38e57
-  data.tar.gz: 13e96259e5668dd4ecfcbeebf5b07320b555d4c55db08aae60a3204e847c974f6efcbfb9a4e29747bbc6f647c23ec73775729152b4b17ca2d79b3c491ea2f7da
+  metadata.gz: c52f3a5bd533a0b00a6d76ee2f2f94f0cfa3bb581823411dfc2fa54d44e81ad807f54ed965b91ea8ee08c37d2861cd890ed501331804c813bab6b57203ccd0c4
+  data.tar.gz: c4c9bf7e8ee84bc89f2d41626de0421773377b1564144e35ce02c0d102b60f2be207ec8b04b5187e7c098691c4b20b9bbc2dff19e4180115ddb442e0712ab8ce

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

@@ -17,22 +17,19 @@ module Uc3DmpId
     MSG_MISSING_PROV_ID = 'No Provenance identifier was provided. \
                            Expected: `{ "dmp_id": { "identifier": "value", "type": "value" }`'
 
+    ORCID_DOMAIN = 'orcid.org'
+    ROR_DOMAIN = 'ror.org'
+
     class << self
       # TODO: Replace this with ElasticSearch
       def search_dmps(args:, logger: nil)
-        client = Uc3DmpDynamo::Client.new
-        return _by_owner(owner_org: args['owner_orcid'], client:, logger:) unless args['owner_orcid'].nil?
+        client = Uc3DmpDynamo::Client.new(table: ENV['DYNAMO_INDEX_TABLE'])
+        return _by_owner(owner: args['owner'], client:, logger:) unless args['owner'].nil?
+        return _by_org(org: args['org'], client:, logger:) unless args['org'].nil?
+        return _by_funder(funder: args['funder'], client:, logger:) unless args['funder'].nil?
+        return _by_featured(client:, logger:) if args.fetch('featured', 'false').to_s.downcase == 'true'
 
-        unless args['owner_org_ror'].nil?
-          return _by_owner_org(owner_org: args['owner_org_ror'], client:,
-                               logger:)
-        end
-        unless args['modification_day'].nil?
-          return _by_mod_day(day: args['modification_day'], client:,
-                             logger:)
-        end
-
-        []
+        return _publicly_visible(client:, logger:)
       end
 
       # Find a DMP based on the contents of the incoming JSON
@@ -134,72 +131,102 @@ module Uc3DmpId
 
       private
 
-      # Fetch the DMP IDs for the specified owner's ORCID (the owner is the :dmphub_owner_id on the DMP ID record)
-      def _by_owner(owner_id:, client: nil, logger: nil)
-        regex = /^([0-9A-Z]{4}-){3}[0-9A-Z]{4}$/
-        raise FinderError, MSG_INVALID_OWNER_ID if owner_id.nil? || (owner_id.to_s =~ regex).nil?
+      # Fetch the DMP IDs for the specified person's ORCID (or email)
+      def _by_owner(owner:, client: nil, logger: nil)
+        orcid_regex = /^([0-9A-Z]{4}-){3}[0-9A-Z]{4}$/
+        email_regex = /^[^\s@]+@[^\s@]+\.[^\s@]+$/
+        email = owner.trim() unless (owner.to_s =~ email_regex).nil?
+
+        if email.nil?
+          args = {
+            filter_expression: 'contains(:people_ids, people)',
+            expression_attribute_values: {
+              ':sk': 'METADATA',
+              ':people_ids': [
+                "http://#{ORCID_DOMAIN}/#{owner}",
+                "https://#{ORCID_DOMAIN}/#{owner}"
+              ]
+            }
+          }
+        else
+          args = {
+            filter_expression: 'contains(:people, people)',
+            expression_attribute_values: { ':sk': 'METADATA', ':people': [email] }
+          }
+        end
+        logger&.debug(message: 'Fetch relevant DMPs _by_owner - scan args', details: args)
+        resp = client.scan(args)
+
+        client = Uc3DmpDynamo::Client.new if client.nil?
+        _process_search_response(response: client.query(args:, logger:))
+      end
+
+      # Fetch the DMP IDs for the specified organization/institution
+      def _by_org(org:, client: nil, logger: nil)
+        regex = /^[a-zA-Z0-9]+$/
+        ror = org.trim() unless (org.to_s =~ regex).nil?
 
         args = {
-          index_name: 'dmphub_owner_id_gsi',
-          key_conditions: {
-            dmphub_owner_id: {
-              attribute_value_list: [
-                "http://orcid.org/#{owner_id}",
-                "https://orcid.org/#{owner_id}"
-              ],
-              comparison_operator: 'IN'
-            }
-          },
-          filter_expression: 'SK = :version',
-          expression_attribute_values: { ':version': Helper::DMP_LATEST_VERSION }
+          filter_expression: 'contains(:affiliation_ids, affiliation_ids)',
+          expression_attribute_values: {
+            ':sk': 'METADATA',
+            ':affiliation_ids': [
+              "http://#{ROR_DOMAIN}/#{ror}",
+              "https://#{ROR_DOMAIN}/#{ror}"
+            ]
+          }
         }
-        logger.info(message: "Querying _by_owner with #{args}") if logger.respond_to?(:info)
+        logger&.debug(message: 'Fetch relevant DMPs _by_org - scan args', details: args)
+        resp = client.scan(args)
+
         client = Uc3DmpDynamo::Client.new if client.nil?
         _process_search_response(response: client.query(args:, logger:))
       end
 
-      # Fetch the DMP IDs for the specified organization/institution (the org is the :dmphub_owner_org
-      # on the DMP ID record)
-      def _by_owner_org(owner_org:, client: nil, logger: nil)
+      # Fetch the DMP IDs for the specified funder
+      def _by_funder(funder:, client: nil, logger: nil)
         regex = /^[a-zA-Z0-9]+$/
-        raise FinderError, MSG_INVALID_OWNER_ID if owner_org.nil? || (owner_org.to_s.downcase =~ regex).nil?
+        ror = funder.trim() unless (funder.to_s =~ regex).nil?
 
         args = {
-          index_name: 'dmphub_owner_org_gsi',
-          key_conditions: {
-            dmphub_owner_org: {
-              attribute_value_list: [
-                "https://ror.org/#{owner_org.to_s.downcase}",
-                "http://ror.org/#{owner_org.to_s.downcase}"
-              ],
-              comparison_operator: 'IN'
-            }
-          },
-          filter_expression: 'SK = :version',
-          expression_attribute_values: { ':version': Helper::DMP_LATEST_VERSION }
+          filter_expression: 'contains(:funder_ids, funder_ids)',
+          expression_attribute_values: {
+            ':sk': 'METADATA',
+            ':funder_ids': [
+              "http://#{ROR_DOMAIN}/#{ror}",
+              "https://#{ROR_DOMAIN}/#{ror}"
+            ]
+          }
         }
-        logger.info(message: "Querying _by_owner_org with #{args}") if logger.respond_to?(:info)
+        logger&.debug(message: 'Fetch relevant DMPs _by_funder - scan args', details: args)
+        resp = client.scan(args)
+
         client = Uc3DmpDynamo::Client.new if client.nil?
         _process_search_response(response: client.query(args:, logger:))
       end
 
-      # Fetch the DMP IDs modified on the specified date (the date is the :dmphub_modification_day on the DMP ID record)
-      def _by_mod_day(day:, client: nil, logger: nil)
-        regex = /^[0-9]{4}(-[0-9]{2}){2}/
-        raise FinderError, MSG_INVALID_OWNER_ID if day.nil? || (day.to_s =~ regex).nil?
+      # Fetch the DMP IDs that are marked as featured
+      def _by_featured(client: nil, logger: nil)
+        args = {
+          filter_expression: ':featured = featured',
+          expression_attribute_values: { ':sk': 'METADATA', ':featured': 1 }
+        }
+        logger&.debug(message: 'Fetch relevant DMPs _by_featured - scan args', details: args)
+        resp = client.scan(args)
+
+        client = Uc3DmpDynamo::Client.new if client.nil?
+        _process_search_response(response: client.query(args:, logger:))
+      end
 
+      # Return all of the publicly visible DMPs
+      def _publicly_visible(client: nil, logger: nil)
         args = {
-          index_name: 'dmphub_modification_day_gsi',
-          key_conditions: {
-            dmphub_modification_day: {
-              attribute_value_list: [day.to_s],
-              comparison_operator: 'IN'
-            }
-          },
-          filter_expression: 'SK = :version',
-          expression_attribute_values: { ':version': Helper::DMP_LATEST_VERSION }
+          filter_expression: ':visibility = visibility',
+          expression_attribute_values: { ':sk': 'METADATA', ':visibility': 'public' }
         }
-        logger.info(message: "Querying _by_mod_day with #{args}") if logger.respond_to?(:info)
+        logger&.debug(message: 'Fetch relevant DMPs _publicly_visible - scan args', details: args)
+        resp = client.scan(args)
+
         client = Uc3DmpDynamo::Client.new if client.nil?
         _process_search_response(response: client.query(args:, logger:))
       end
@@ -212,8 +239,8 @@ module Uc3DmpId
           next if item.nil?
 
           dmp = item['dmp'].nil? ? JSON.parse({ dmp: item }.to_json) : item
-          dmp = _remove_narrative_if_private(json: dmp)
-          Helper.cleanse_dmp_json(json: dmp)
+          # dmp = _remove_narrative_if_private(json: dmp)
+          # Helper.cleanse_dmp_json(json: dmp)
         end
         results.compact.uniq
       end

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

@@ -204,23 +204,13 @@ module Uc3DmpId
         # The `dmphub_modifications` array will ONLY ever have things the harvester mods know about
         # so just find them and update the status accordingly
         mods = resp.dup
-
-puts 'MODIFICATIONS RECORD'
-puts mods
-
-puts 'INCOMING RECORD'
-puts json['dmphub_modifications']
-
-
         json['dmphub_modifications'].each do |entry|
           next if entry.fetch('dmproadmap_related_identifiers', []).empty?
 
           entry['dmproadmap_related_identifiers'].each do |related|
-            related_id = mods['related_works'][related.identifier] if related.respond_to?(:identifier)
-            related_id = mods['related_works'][related['identifier']] if related_id.nil?
-            next if related_id.nil?
+            next if mods['related_works'][related.identifier].nil?
 
-            mods['related_works'][related_id]['status'] = related['status']
+            mods['related_works'][related.identifier]['status'] = related['status']
           end
         end
 

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

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Uc3DmpId
-  VERSION = '0.1.50'
+  VERSION = '0.1.52'
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: uc3-dmp-id
 version: !ruby/object:Gem::Version
-  version: 0.1.50
+  version: 0.1.52
 platform: ruby
 authors:
 - Brian Riley
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2024-07-08 00:00:00.000000000 Z
+date: 2024-07-09 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: json