adiwg-mdtranslator 2.19.0.pre.beta.25 → 2.20.0.pre.beta.3

data/lib/adiwg/mdtranslator/writers/dcat_us/readme.md

@@ -0,0 +1,10 @@
+
+## dcat_us
+
+### Supported versions
+
+> 0.0.x (dcat_us is not currently versioned)
+
+### Writer for Data Catalog Vocabulary (DCAT) v1.1
+
+

data/lib/adiwg/mdtranslator/writers/dcat_us/sections/dcat_us_access_level.rb

@@ -0,0 +1,55 @@
+require 'jbuilder'
+
+module ADIWG
+   module Mdtranslator
+      module Writers
+         module Dcat_us
+            module AccessLevel
+
+               def self.build(intObj)
+
+                  publicArray = ['unclassified', 'unrestricted', 'licenseUnrestricted', 'licenseEndUser']
+                  nonPublicArray = ['restricted','confidential','secret','topSecret','forOfficialUseOnly','protected','intellectualPropertyRights','restricted','otherRestrictions','private','statutory','confidential','traditionalKnowledge','personallyIdentifiableInformation']
+                  restrictedPublicArray = ['sensitiveButUnclassified','limitedDistribution','copyright','patent','patentPending','trademark','license','licenseDistributor','in-confidence','threatenedOrEndangered']
+
+                  resourceInfo = intObj[:metadata][:resourceInfo]
+                  legalConstraints = resourceInfo[:constraints]&.select { |constraint| constraint[:type] == 'legal' }
+                  securityConstraints = resourceInfo[:constraints]&.select { |constraint| constraint[:type] == 'security' }
+
+                  accessLevelCodes = []
+
+                  # Gather codes from security constraints and legal constraints
+                  unless securityConstraints.empty?
+                     securityConstraints.each do |securityConstraint|
+                        code = securityConstraint[:securityConstraint][:classCode]
+                        accessLevelCodes << code                        
+                     end
+                  end
+                  unless legalConstraints.empty?
+                     legalConstraints.each do |legalConstraint|
+                        codes = legalConstraint.dig(:legalConstraint, :accessCodes)
+                        accessLevelCodes.push(*codes)
+                     end
+                  end
+
+                  # return access level that is most restrictive
+                  accessLevelCodes.uniq.each do |code|
+                     if nonPublicArray.include? code
+                        return 'non-public'
+                     end
+                  end
+                  accessLevelCodes.uniq.each do |code|
+                     if restrictedPublicArray.include? code
+                        return 'restricted public'
+                     end
+                  end
+                  
+                  return 'public'
+               end           
+
+            end
+         end
+      end
+   end
+end
+

data/lib/adiwg/mdtranslator/writers/dcat_us/sections/dcat_us_access_url.rb

@@ -0,0 +1,17 @@
+require 'jbuilder'
+
+module ADIWG
+   module Mdtranslator
+      module Writers
+         module Dcat_us
+            module AccessURL
+
+               def self.build(option)
+                  option[:olResURI] if option[:olResURI].end_with?('.html')
+                end                
+
+            end
+         end
+      end
+   end
+end

data/lib/adiwg/mdtranslator/writers/dcat_us/sections/dcat_us_accrualPeriodicity.rb

@@ -0,0 +1,45 @@
+require 'jbuilder'
+
+module ADIWG
+   module Mdtranslator
+      module Writers
+         module Dcat_us
+            module AccrualPeriodicity
+
+               def self.build(intObj)
+
+                  frequency_mapping = {
+                     'decennial' => 'R/P10Y',
+                     'quadrennial' => 'R/P4Y',
+                     'annual' => 'R/P1Y',
+                     'bimonthly' => 'R/P2M or R/P0.5M',
+                     'semiweekly' => 'R/P3.5D',
+                     'daily' => 'R/P1D',
+                     'biweekly' => 'R/P2W or R/P0.5W',
+                     'semiannual' => 'R/P6M',
+                     'biennial' => 'R/P2Y',
+                     'triennial' => 'R/P3Y',
+                     'three times a week' => 'R/P0.33W',
+                     'three times a month' => 'R/P0.33M',
+                     'continuously updated' => 'R/PT1S',
+                     'monthly' => 'R/P1M',
+                     'quarterly' => 'R/P3M',
+                     'semimonthly' => 'R/P0.5M',
+                     'three times a year' => 'R/P4M',
+                     'weekly' => 'R/P1W',
+                     'hourly' => 'R/PT1H'
+                  }
+
+                  frequency = intObj[:metadata][:metadataInfo][:metadataMaintenance][:frequency]
+                  
+                  unless frequency.nil?
+                     frequency_code = frequency_mapping[frequency.downcase]
+                  end
+                  return frequency_code
+                end                
+
+            end
+         end
+      end
+   end
+end

data/lib/adiwg/mdtranslator/writers/dcat_us/sections/dcat_us_bureau_code.rb

@@ -0,0 +1,33 @@
+require 'jbuilder'
+
+module ADIWG
+   module Mdtranslator
+      module Writers
+         module Dcat_us
+            module BureauCode
+
+               def self.build(intObj)
+                
+                    responsibleParties = intObj[:metadata][:resourceInfo][:citation][:responsibleParties]
+                    contacts = []
+                    responsibleParties.each do |party|
+                        contactId = party[:parties][0][:contactId]
+                        contacts << Dcat_us.get_contact_by_id(contactId)
+                    end
+
+                    bureauContacts = contacts&.select { |contact| contact[:externalIdentifier].any? { |id| id[:namespace] == 'bureauCode'} }
+
+                    bureauCodes = []
+                    bureauContacts.each do |contact|
+                        bureauCode = contact[:externalIdentifier].find { |id| id[:namespace] == 'bureauCode' }
+                        bureauCodes << bureauCode[:identifier]
+                    end
+
+                    return bureauCodes
+                end
+                
+            end
+         end
+      end
+   end
+end

data/lib/adiwg/mdtranslator/writers/dcat_us/sections/dcat_us_contact_point.rb

@@ -0,0 +1,29 @@
+require 'jbuilder'
+
+module ADIWG
+   module Mdtranslator
+      module Writers
+         module Dcat_us
+            module ContactPoint
+
+               def self.build(intObj)
+                  resourceInfo = intObj[:metadata][:resourceInfo]
+                  pointOfContact = resourceInfo[:pointOfContacts][0]
+                  contactId = pointOfContact[:parties][0][:contactId]
+
+                  contact = Dcat_us.get_contact_by_id(contactId)
+                  fn = contact[:name]
+                  hasEmail = contact[:eMailList][0]
+
+                  Jbuilder.new do |json|
+                     json.set!('@type', 'vcard:Contact')
+                     json.set!('fn', fn)
+                     json.set!('hasEmail', hasEmail)
+                  end
+
+               end
+            end
+         end
+      end
+   end
+end

data/lib/adiwg/mdtranslator/writers/dcat_us/sections/dcat_us_dcat_us.rb

@@ -0,0 +1,104 @@
+require 'jbuilder'
+require_relative 'dcat_us_keyword'
+require_relative 'dcat_us_publisher'
+require_relative 'dcat_us_contact_point'
+require_relative 'dcat_us_identifier'
+require_relative 'dcat_us_distribution'
+require_relative 'dcat_us_spatial'
+require_relative 'dcat_us_temporal'
+require_relative 'dcat_us_modified'
+require_relative 'dcat_us_access_level'
+require_relative 'dcat_us_rights'
+require_relative 'dcat_us_license'
+require_relative 'dcat_us_issued'
+require_relative 'dcat_us_described_by'
+require_relative 'dcat_us_is_part_of'
+require_relative 'dcat_us_theme'
+require_relative 'dcat_us_references'
+require_relative 'dcat_us_landing_page'
+require_relative 'dcat_us_system_of_records'
+require_relative 'dcat_us_description'
+require_relative 'dcat_us_described_by_type'
+require_relative 'dcat_us_accrualPeriodicity'
+require_relative 'dcat_us_language'
+require_relative 'dcat_us_primaryITInvestmentUII'
+require_relative 'dcat_us_program_code'
+require_relative 'dcat_us_bureau_code'
+
+module ADIWG
+   module Mdtranslator
+      module Writers
+         module Dcat_us
+
+            def self.build(intObj, responseObj)
+               metadataInfo = intObj[:metadata][:metadataInfo]
+               resourceInfo = intObj[:metadata][:resourceInfo]
+               citation = resourceInfo[:citation]
+
+               title = citation[:title]
+               description = Description.build(intObj)
+               keyword = Keyword.build(intObj)
+               modified = Modified.build(intObj)
+               publisher = Publisher.build(intObj)
+               contactPoint = ContactPoint.build(intObj)
+               accessLevel = AccessLevel.build(intObj)
+               identifier = Identifier.build(intObj)
+               distribution = Distribution.build(intObj)
+               rights = Rights.build(intObj, accessLevel)
+               spatial = Spatial.build(intObj)
+               temporal = Temporal.build(intObj)
+               license = License.build(intObj)
+               issued = Issued.build(intObj)
+               language = Language.build(intObj)
+               describedBy = DescribedBy.build(intObj)
+               isPartOf = IsPartOf.build(intObj)
+               theme = Theme.build(intObj)
+               references = References.build(intObj)
+               landingPage = LandingPage.build(intObj)
+               systemOfRecords = SystemOfRecords.build(intObj)
+               describedByType = DescribedByType.build(intObj)
+               accrualPeriodicity = AccrualPeriodicity.build(intObj)
+               primaryITInvestmentUII = PrimaryITInvestmentUII.build(intObj)
+               programCode = ProgramCode.build(intObj)
+               bureauCode = BureauCode.build(intObj)
+
+               @Namespace = ADIWG::Mdtranslator::Writers::Dcat_us
+
+               Jbuilder.new do |json|
+                  json.set!('@type', 'dcat:Dataset')
+                  json.set!('title', title)
+                  json.set!('description', description)
+                  json.set!('keyword', keyword)
+                  json.set!('modified', modified)
+                  json.set!('publisher', publisher) 
+                  json.set!('contactPoint', contactPoint)
+                  json.set!('identifier', identifier)
+                  json.set!('accessLevel', accessLevel)
+                  json.set!('bureauCode', bureauCode)
+                  json.set!('programCode', programCode)
+                  json.set!('distribution', distribution)
+
+                  json.set!('license', license)
+                  json.set!('rights', rights)
+                  json.set!('spatial', spatial)
+                  json.set!('temporal', temporal)
+
+                  json.set!('issued', issued)
+                  json.set!('accrualPeriodicity', accrualPeriodicity)
+                  json.set!('language', language)
+                  # json.set!('dataQuality', metadataInfo[:metadataMaintenance][:maintenanceNote])
+                  json.set!('theme', theme)
+                  json.set!('references', references)
+                  json.set!('landingPage', landingPage)
+                  json.set!('isPartOf', isPartOf)
+                  json.set!('systemOfRecords', systemOfRecords)
+                  json.set!('primaryITInvestmentUII', primaryITInvestmentUII)
+                  json.set!('describedBy', describedBy)
+                  json.set!('describedByType', describedByType)
+                  # json.set!('conformsTo', metadataInfo[:metadataStandards][0][:standardName])
+               end
+            end
+         end
+      end
+   end
+end

data/lib/adiwg/mdtranslator/writers/dcat_us/sections/dcat_us_described_by.rb

@@ -0,0 +1,29 @@
+require 'jbuilder'
+
+module ADIWG
+   module Mdtranslator
+      module Writers
+         module Dcat_us
+            module DescribedBy
+
+               def self.build(intObj)
+                  dataDictionaries = intObj[:dataDictionaries]
+                  dataDictionaries.each do |dataDictionary|
+                     unless dataDictionary[:includedWithDataset]
+                        onlineResources = dataDictionary[:citation][:onlineResources]
+                        onlineResources.each do |resource|
+                           if resource[:olResURI]
+                              return resource[:olResURI]
+                           end
+                        end
+                     end
+                  end
+                
+                  nil
+                end                
+                
+            end
+         end
+      end
+   end
+end

data/lib/adiwg/mdtranslator/writers/dcat_us/sections/dcat_us_described_by_type.rb

@@ -0,0 +1,33 @@
+require 'jbuilder'
+
+module ADIWG
+   module Mdtranslator
+      module Writers
+         module Dcat_us
+            module DescribedByType
+
+               def self.build(intObj)
+
+                  #  metadataInfo[:metadataOnlineOptions][0][:olResProtocol]
+                  dataDictionaries = intObj[:dataDictionaries]
+                  describedByType = ''
+                  dataDictionaries.each do |dataDictionary|
+                     unless dataDictionary[:includedWithDataset]
+                        onlineResources = dataDictionary[:citation][:onlineResources]
+                        onlineResources.each do |resource|
+                           if resource[:olResURI] && !resource[:olResURI].end_with?('.html')
+                              describedByType = resource[:olResProtocol]
+                              break
+                           end
+                        end
+                     end
+                  end
+                
+                  describedByType
+               end                
+
+            end
+         end
+      end
+   end
+end

data/lib/adiwg/mdtranslator/writers/dcat_us/sections/dcat_us_description.rb

@@ -0,0 +1,19 @@
+module ADIWG
+   module Mdtranslator
+      module Writers
+         module Dcat_us
+            module Description
+
+               def self.build(intObj)
+                  resourceInfo = intObj.dig(:metadata, :resourceInfo)
+                  description = resourceInfo&.dig(:abstract)
+
+                  return nil unless description
+               end
+                                    
+
+            end
+         end
+      end
+   end
+end

data/lib/adiwg/mdtranslator/writers/dcat_us/sections/dcat_us_distribution.rb

@@ -0,0 +1,59 @@
+require 'jbuilder'
+require_relative 'dcat_us_access_url'
+require_relative 'dcat_us_download_url'
+require_relative 'dcat_us_media_type'
+
+
+module ADIWG
+  module Mdtranslator
+    module Writers
+      module Dcat_us
+        module Distribution
+
+          def self.build(intObj)
+
+            resourceDistributions = intObj.dig(:metadata, :distributorInfo)
+            distributions = []
+
+            resourceDistributions&.each do |resource|
+              description = resource[:description] || ''
+              break_flag = false  # Flag to control nested loop breaks
+
+              resource[:distributor]&.each do |distributor|
+                break if break_flag
+
+                distributor[:transferOptions]&.each do |transfer|
+                  break if break_flag
+
+                  mediaType = MediaType.build(transfer)
+
+                  transfer[:onlineOptions]&.each do |option|
+                    next unless option[:olResURI]
+                    accessURL = AccessURL.build(option)
+                    downloadURL = DownloadURL.build(option)
+                    title = option[:olResName] || ''
+
+                    distribution = Jbuilder.new do |json|
+                      json.set!('@type', 'dcat:Distribution')
+                      json.set!('description', description)
+                      json.set!('accessURL', accessURL) if accessURL
+                      json.set!('downloadURL', downloadURL) if downloadURL
+                      json.set!('mediaType', mediaType)
+                      json.set!('title', title)
+                    end
+
+                    distributions << distribution.attributes!
+                    break_flag = true
+                    break
+                  end
+                end
+              end
+            end
+            distributions
+          end
+
+        end
+      end
+    end
+  end
+end

data/lib/adiwg/mdtranslator/writers/dcat_us/sections/dcat_us_download_url.rb

@@ -0,0 +1,17 @@
+require 'jbuilder'
+
+module ADIWG
+   module Mdtranslator
+      module Writers
+         module Dcat_us
+            module DownloadURL
+
+               def self.build(option)
+                  option[:olResURI] unless option[:olResURI].end_with?('.html')
+                end                
+
+            end
+         end
+      end
+   end
+end

data/lib/adiwg/mdtranslator/writers/dcat_us/sections/dcat_us_identifier.rb

@@ -0,0 +1,29 @@
+module ADIWG
+   module Mdtranslator
+      module Writers
+         module Dcat_us
+            module Identifier
+
+               def self.build(intObj)
+                  citation = intObj.dig(:metadata, :resourceInfo, :citation)
+                  identifiers = citation&.dig(:identifiers)
+                  onlineResources = citation&.dig(:onlineResources)
+                  uri = onlineResources.dig(0, :olResURI)
+                
+                  namespace_is_doi = identifiers&.any? { |identifier| identifier[:namespace]&.casecmp?("DOI") }
+                
+                  if namespace_is_doi
+                    return uri
+                  elsif uri && uri.downcase.include?("doi")
+                    return uri
+                  end
+                
+                  nil
+               end
+                                    
+
+            end
+         end
+      end
+   end
+end

data/lib/adiwg/mdtranslator/writers/dcat_us/sections/dcat_us_is_part_of.rb

@@ -0,0 +1,27 @@
+require 'jbuilder'
+
+module ADIWG
+   module Mdtranslator
+      module Writers
+         module Dcat_us
+            module IsPartOf
+
+               def self.build(intObj)
+                  associatedResources = intObj.dig(:metadata, :associatedResources)
+                
+                  associatedResources.each do |resource|
+                    next unless resource[:initiativeType] == "collection" && resource[:associationType] == "collectiveTitle"
+                
+                    onlineResources = resource.dig(:resourceCitation, :onlineResources) || []
+                    uri = onlineResources.find { |onlineResource| onlineResource[:olResURI] }&.dig(:olResURI)
+                    return uri if uri
+                  end
+                
+                  nil
+                end                
+
+            end
+         end
+      end
+   end
+end

data/lib/adiwg/mdtranslator/writers/dcat_us/sections/dcat_us_issued.rb

@@ -0,0 +1,19 @@
+require 'jbuilder'
+
+module ADIWG
+   module Mdtranslator
+      module Writers
+         module Dcat_us
+            module Issued
+
+               def self.build(intObj)
+                  dates = intObj[:metadata][:resourceInfo][:citation][:dates].map { |obj| obj[:date] }
+                  earliest_date = dates.min
+                  return earliest_date
+                end
+                
+            end
+         end
+      end
+   end
+end

data/lib/adiwg/mdtranslator/writers/dcat_us/sections/dcat_us_keyword.rb

@@ -0,0 +1,19 @@
+module ADIWG
+   module Mdtranslator
+      module Writers
+         module Dcat_us
+            module Keyword
+ 
+               def self.build(intObj)
+                  resourceInfo = intObj.dig(:metadata, :resourceInfo)
+                  keywords = resourceInfo&.dig(:keywords)
+                
+                  return keywords&.flat_map { |keyword| keyword[:keywords]&.map { |kw| kw[:keyword] } } || []
+                end                
+                
+            end
+         end
+      end
+   end
+end
+ 

data/lib/adiwg/mdtranslator/writers/dcat_us/sections/dcat_us_landing_page.rb

@@ -0,0 +1,27 @@
+require 'jbuilder'
+
+module ADIWG
+   module Mdtranslator
+      module Writers
+         module Dcat_us
+            module LandingPage
+
+               def self.build(intObj)
+                  onlineResources = intObj.dig(:metadata, :resourceInfo, :citation, :onlineResources)
+                
+                  if onlineResources
+                    onlineResources.each do |resource|
+                      if resource.dig(:olResFunction) == 'landingPage'
+                        return resource.dig(:olResURI)
+                      end
+                    end
+                  end
+                  
+                  return nil
+                end                
+                
+            end
+         end
+      end
+   end
+end

data/lib/adiwg/mdtranslator/writers/dcat_us/sections/dcat_us_language.rb

@@ -0,0 +1,23 @@
+module ADIWG
+   module Mdtranslator
+      module Writers
+         module Dcat_us
+            module Language
+ 
+               def self.build(intObj)
+                  metadataInfo = intObj.dig(:metadata, :metadataInfo)
+                  defaultLanguage = metadataInfo&.dig(:defaultMetadataLocale, :languageCode)
+                  additionalLanguage = metadataInfo&.dig(:otherMetadataLocales)
+                  languages = []
+                  languages << defaultLanguage if defaultLanguage
+                  additionalLanguage.each { |lang| languages << lang[:languageCode] } if additionalLanguage
+
+                  return languages
+                end                
+                
+            end
+         end
+      end
+   end
+end
+ 

data/lib/adiwg/mdtranslator/writers/dcat_us/sections/dcat_us_license.rb

@@ -0,0 +1,18 @@
+module ADIWG
+   module Mdtranslator
+      module Writers
+         module Dcat_us
+            module License
+
+               def self.build(intObj)
+                  resourceInfo = intObj.dig(:metadata, :resourceInfo)
+                  title = resourceInfo.dig(:constraints, 0, :reference, 0, :title)
+                  license = title || 'https://creativecommons.org/publicdomain/zero/1.0/'
+                  license
+               end
+
+            end
+         end
+      end
+   end
+end

data/lib/adiwg/mdtranslator/writers/dcat_us/sections/dcat_us_media_type.rb

@@ -0,0 +1,17 @@
+require 'jbuilder'
+
+module ADIWG
+   module Mdtranslator
+      module Writers
+         module Dcat_us
+            module MediaType
+
+               def self.build(transfer)
+                  transfer.dig(:distributionFormats)&.find { |format| format.dig(:formatSpecification, :title) }&.dig(:formatSpecification, :title) || ''
+                end                
+
+            end
+         end
+      end
+   end
+end