uc3-dmp-id 0.1.58 → 0.1.60

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: '0987f59845998c2a2e3409986fe93f8b176f0f3ad175f2b25384fd1c558dd943'
-  data.tar.gz: 7ee6181e791f32a788c96e4f5c4a174cb510f7b89a99d2bd0dc4e575e457efe8
+  metadata.gz: 791179155f0be964f0cfc186f26de88e9f0774ee9ebb671f7856074b1d40639e
+  data.tar.gz: 0bf38f90d0863863362ea4fb5357566d8b262c2036840d78519796e2d55d893a
 SHA512:
-  metadata.gz: a66993ced9e0c61fac91fb754f90f637c7eaf630c8751b3dd2ee7b4b396f3741e71ec48fd64e387c508cc68b2bcad4f9cf89dda747a56b2d0eb2fb42720a6c30
-  data.tar.gz: 42ca4b60f5e094acd492257dd97da3fc1191912c9f1dbad4fbbe614c7ad34ea02db617ea1e9501e02ede5a5a4e4493d5e7041d365bceff6bbc7f6d604cf34ad9
+  metadata.gz: e938eb3ddaa088e2cc17bb6a4202de8673c5eeb046c47207e1c040eff8775fdede2b835a90758ab8e892490677b93502c205b0244fae3884299498caf27575a8
+  data.tar.gz: ae45834944f078ab9e0d3fae1c6b32d2f5015c5b74b46c526c62b457b22e329662c6049ba4dd090e67f4c7f4208d9a1f858cb62b4dc52fa8b1e2bc1b4b17f692

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

@@ -17,19 +17,32 @@ 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'
+    ORCID_DOMAIN = 'https://orcid.org/'
+    ROR_DOMAIN = 'https://ror.org/'
 
     class << self
       # TODO: Replace this with ElasticSearch
       def search_dmps(args:, logger: nil)
+        # Fetch the DMPs for each of the possible filter options
         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'
+        owner_pks = _by_owner(owner: args['owner'], client:, logger:) unless args['owner'].nil?
+        org_pks = _by_org(org: args['org'], client:, logger:) unless args['org'].nil?
+        funder_pks = _by_funder(funder: args['funder'], client:, logger:) unless args['funder'].nil?
+        return [] if owner_pks.empty? && org_pks.empty? && funder_pks.empty?
 
-        return _publicly_visible(client:, logger:)
+        # Only use the DMPs that fit all of the filter criteria
+        dmps = owner_pks & org_pks & funder_pks
+        return [] if dmps.nil? || dmps.empty?
+
+        # Perfom search operations
+
+        # Handle sort
+
+        # Handle pagination
+
+        # Fetch full DMP records
+        client = Uc3DmpDynamo::Client.new(table: ENV['DYNAMO_TABLE'])
+        _fetch_dmps(client:, pks: dmps.map { |dmp| dmp['pk'] }, logger:)
       end
 
       # Find a DMP based on the contents of the incoming JSON
@@ -133,79 +146,47 @@ module Uc3DmpId
 
       # 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}$/
+        orcid_regex = /^([0-9a-zA-Z]{4}-){3}[0-9a-zA-Z]{4}$/
         email_regex = /^[^\s@]+@[^\s@]+\.[^\s@]+$/
-        email = owner.strip unless (owner.to_s.strip.gsub('%40', '@') =~ email_regex).nil?
-
-        if !(owner.to_s.strip =~ orcid_regex).nil?
-          args = {
-            filter_expression: 'contains(:people_ids, people) AND :sk = sk',
-            expression_attribute_values: {
-              ':sk': 'METADATA',
-              ':people_ids': [
-                "http://#{ORCID_DOMAIN}/#{owner}",
-                "https://#{ORCID_DOMAIN}/#{owner}"
-              ]
-            }
-          }
-        elsif !email.nil?
-          args = {
-            filter_expression: 'contains(:people, people) AND :sk = sk',
-            expression_attribute_values: { ':sk': 'METADATA', ':people': [email.gsub('%40', '@')] }
-          }
-        else
-          # It wasn't an email or ORCID, so return an empty array
-          return []
-        end
-        logger&.debug(message: 'Fetch relevant DMPs _by_owner - scan args', details: args)
-        client = Uc3DmpDynamo::Client.new if client.nil?
-        _process_search_response(response: client.scan(args:))
+        orcid = owner.to_s.strip
+        return [] if (orcid =~ orcid_regex).nil? && (orcid =~ email_regex).nil?
+
+        orcid = "#{ORCID_DOMAIN}#{orcid}" unless (orcid =~ orcid_regex).nil?
+        resp = client.get_item(key: { PK: 'PERSON_INDEX', SK: orcid }, logger:)
+        return [] unless resp.is_a?(Hash)
+
+        logger&.debug(message: "DMPs for PERSON #{orcid}", details: resp)
+        resp.fetch('dmps', [])
       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.strip unless (org.to_s =~ regex).nil?
+        ror = "#{ROR_DOMAIN}/#{org.strip}" unless (org.to_s =~ regex).nil?
 
-        args = {
-          filter_expression: 'contains(:affiliation_ids, affiliation_ids) AND :sk = sk',
-          expression_attribute_values: {
-            ':sk': 'METADATA',
-            ':affiliation_ids': [
-              "http://#{ROR_DOMAIN}/#{ror}",
-              "https://#{ROR_DOMAIN}/#{ror}"
-            ]
-          }
-        }
-        logger&.debug(message: 'Fetch relevant DMPs _by_org - scan args', details: args)
-        client = Uc3DmpDynamo::Client.new if client.nil?
-        _process_search_response(response: client.scan(args:))
+        resp = client.get_item(key: { PK: 'AFFILIATION_INDEX', SK: ror }, logger:)
+        return [] unless resp.is_a?(Hash)
+
+        logger&.debug(message: "DMPs for AFFILIATION #{ror}", details: resp)
+        resp.fetch('dmps', [])
       end
 
       # Fetch the DMP IDs for the specified funder
       def _by_funder(funder:, client: nil, logger: nil)
         regex = /^[a-zA-Z0-9]+$/
-        ror = funder.strip unless (funder.to_s =~ regex).nil?
+        ror = "#{ROR_DOMAIN}/#{funder.strip}" unless (funder.to_s =~ regex).nil?
 
-        args = {
-          filter_expression: 'contains(:funder_ids, funder_ids) AND :sk = sk',
-          expression_attribute_values: {
-            ':sk': 'METADATA',
-            ':funder_ids': [
-              "http://#{ROR_DOMAIN}/#{ror}",
-              "https://#{ROR_DOMAIN}/#{ror}"
-            ]
-          }
-        }
-        logger&.debug(message: 'Fetch relevant DMPs _by_funder - scan args', details: args)
-        client = Uc3DmpDynamo::Client.new if client.nil?
-        _process_search_response(response: client.scan(args:))
+        resp = client.get_item(key: { PK: 'FUNDER_INDEX', SK: ror }, logger:)
+        return [] unless resp.is_a?(Hash)
+
+        logger&.debug(message: "DMPs for FUNDER #{ror}", details: resp)
+        resp.fetch('dmps', [])
       end
 
       # Fetch the DMP IDs that are marked as featured
       def _by_featured(client: nil, logger: nil)
         args = {
-          filter_expression: ':featured = featured AND :sk = sk',
+          filter_expression: 'featured = :featured AND SK = :sk',
           expression_attribute_values: { ':sk': 'METADATA', ':featured': 1 }
         }
         logger&.debug(message: 'Fetch relevant DMPs _by_featured - scan args', details: args)
@@ -216,7 +197,7 @@ module Uc3DmpId
       # Return all of the publicly visible DMPs
       def _publicly_visible(client: nil, logger: nil)
         args = {
-          filter_expression: ':visibility = visibility AND :sk = sk',
+          filter_expression: 'visibility = :visibility AND SK = :sk',
           expression_attribute_values: { ':sk': 'METADATA', ':visibility': 'public' }
         }
         logger&.debug(message: 'Fetch relevant DMPs _publicly_visible - scan args', details: args)
@@ -224,6 +205,11 @@ module Uc3DmpId
         _process_search_response(response: client.scan(args:))
       end
 
+      # Fetches all of the DMPs by their PKs
+      def _fetch_dmps(client:, pks:, logger: null)
+
+      end
+
       # Transform the search results so that we do not include any of the DMPHub specific metadata
       def _process_search_response(response:)
         return [] unless response.is_a?(Array) && response.any?

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

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Uc3DmpId
-  VERSION = '0.1.58'
+  VERSION = '0.1.60'
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: uc3-dmp-id
 version: !ruby/object:Gem::Version
-  version: 0.1.58
+  version: 0.1.60
 platform: ruby
 authors:
 - Brian Riley
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2024-07-09 00:00:00.000000000 Z
+date: 2024-07-10 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: json