datacite-mapping 0.1.0

data/examples/writing.rb

@@ -0,0 +1,49 @@
+#!/usr/bin/env ruby
+
+require 'datacite/mapping'
+include Datacite::Mapping
+
+# Based on "Example for a simple dataset"
+# http://schema.datacite.org/meta/kernel-3/example/datacite-example-dataset-v3.0.xml
+resource = Resource.new(
+  identifier: Identifier.new(value: '10.5072/D3P26Q35R-Test'),
+  creators: [
+    Creator.new(name: 'Fosmire, Michael'),
+    Creator.new(name: 'Wertz, Ruth'),
+    Creator.new(name: 'Purzer, Senay')
+  ],
+  titles: [
+    Title.new(value: 'Critical Engineering Literacy Test (CELT)')
+  ],
+  publisher: 'Purdue University Research Repository (PURR)',
+  publication_year: 2013,
+  subjects: [
+    Subject.new(value: 'Assessment'),
+    Subject.new(value: 'Information Literacy'),
+    Subject.new(value: 'Engineering'),
+    Subject.new(value: 'Undergraduate Students'),
+    Subject.new(value: 'CELT'),
+    Subject.new(value: 'Purdue University')
+  ],
+  language: 'en',
+  resource_type: ResourceType.new(resource_type_general: ResourceTypeGeneral::DATASET, value: 'Dataset'),
+  version: '1',
+  descriptions: [
+    Description.new(
+      type: DescriptionType::ABSTRACT,
+      value: 'We developed an instrument, Critical Engineering Literacy Test (CELT), which is a multiple choice instrument
+              designed to measure undergraduate students’ scientific and information literacy skills. It requires students to
+              first read a technical memo and, based on the memo’s arguments, answer eight multiple choice and six open-ended
+              response questions. We collected data from 143 first-year engineering students and conducted an item analysis. The
+              KR-20 reliability of the instrument was .39. Item difficulties ranged between .17 to .83. The results indicate low
+              reliability index but acceptable levels of item difficulties and item discrimination indices. Students were most
+              challenged when answering items measuring scientific and mathematical literacy (i.e., identifying incorrect
+              information).'
+    )
+  ]
+)
+
+xml = resource.save_to_xml
+formatter = REXML::Formatters::Pretty.new
+formatter.compact = true
+formatter.write(xml, $stdout)

data/lib/datacite/mapping.rb

@@ -0,0 +1,36 @@
+# Module for working with the [DataCite metadata schema](https://schema.datacite.org/meta/kernel-3/index.html)
+module Datacite
+  # Maps DataCite XML to Ruby objects
+  module Mapping
+
+    Dir.glob(File.expand_path('../mapping/*.rb', __FILE__)).sort.each(&method(:require))
+
+    class << self
+      attr_writer :log
+    end
+
+    # Gets the logger for the module. Default logger logs to `$stdout`.
+    # @return [Logger] the logger
+    def self.log
+      self.log_device = $stdout unless @log
+      @log
+    end
+
+    # Sets the log device. Defaults to `$stdout`
+    # @param value [IO] the log device
+    def self.log_device=(value)
+      @log = new_logger(logdev: value)
+    end
+
+    private
+
+    def self.new_logger(logdev:, level: Logger::DEBUG, shift_age: 10, shift_size: 1024 * 1024)
+      logger = Logger.new(logdev, shift_age, shift_size)
+      logger.level = level
+      logger.formatter = proc do |severity, datetime, progname, msg|
+        "#{datetime.to_time.utc} #{severity} -#{progname}- #{msg}\n"
+      end
+      logger
+    end
+  end
+end

data/lib/datacite/mapping/alternate_identifier.rb

@@ -0,0 +1,45 @@
+require 'xml/mapping_extensions'
+
+module Datacite
+  module Mapping
+
+    # An identifier or identifiers other than the primary {Identifier}
+    # applied to the {Resource}.
+    class AlternateIdentifier
+      include XML::Mapping
+
+      root_element_name 'alternateIdentifier'
+
+      text_node :type, '@alternateIdentifierType'
+      text_node :value, 'text()'
+
+      alias_method :_type=, :type=
+      private :_type=
+      alias_method :_value=, :value=
+      private :_value=
+
+      # Initializes a new {AlternateIdentifier}
+      # @param type [String] the identifier type
+      # @param value [String] the identifier value
+      def initialize(type:, value:)
+        self.type = type
+        self.value = value
+      end
+
+      # Sets the type. Cannot be nil.
+      # @param val [String] the identifier type
+      def type=(val)
+        fail ArgumentError, 'No identifier type provided' unless val
+        self._type = val
+      end
+
+      # Sets the value. Cannot be nil.
+      # @param val [String] the value
+      def value=(val)
+        fail ArgumentError, 'No identifier value provided' unless val
+        self._value = val
+      end
+
+    end
+  end
+end

data/lib/datacite/mapping/contributor.rb

@@ -0,0 +1,125 @@
+require 'xml/mapping_extensions'
+require_relative 'name_identifier'
+
+module Datacite
+  module Mapping
+
+    # Controlled vocabulary of contributor types.
+    class ContributorType < TypesafeEnum::Base
+      # @!parse CONTACT_PERSON = ContactPerson
+      new :CONTACT_PERSON, 'ContactPerson'
+
+      # @!parse DATA_COLLECTOR = DataCollector
+      new :DATA_COLLECTOR, 'DataCollector'
+
+      # @!parse DATA_CURATOR = DataCurator
+      new :DATA_CURATOR, 'DataCurator'
+
+      # @!parse DATA_MANAGER = DataManager
+      new :DATA_MANAGER, 'DataManager'
+
+      # @!parse DISTRIBUTOR = Distributor
+      new :DISTRIBUTOR, 'Distributor'
+
+      # @!parse EDITOR = Editor
+      new :EDITOR, 'Editor'
+
+      # @!parse FUNDER = Funder
+      new :FUNDER, 'Funder'
+
+      # @!parse HOSTING_INSTITUTION = HostingInstitution
+      new :HOSTING_INSTITUTION, 'HostingInstitution'
+
+      # @!parse PRODUCER = Producer
+      new :PRODUCER, 'Producer'
+
+      # @!parse PROJECT_LEADER = ProjectLeader
+      new :PROJECT_LEADER, 'ProjectLeader'
+
+      # @!parse PROJECT_MANAGER = ProjectManager
+      new :PROJECT_MANAGER, 'ProjectManager'
+
+      # @!parse PROJECT_MEMBER = ProjectMember
+      new :PROJECT_MEMBER, 'ProjectMember'
+
+      # @!parse REGISTRATION_AGENCY = RegistrationAgency
+      new :REGISTRATION_AGENCY, 'RegistrationAgency'
+
+      # @!parse REGISTRATION_AUTHORITY = RegistrationAuthority
+      new :REGISTRATION_AUTHORITY, 'RegistrationAuthority'
+
+      # @!parse RELATED_PERSON = RelatedPerson
+      new :RELATED_PERSON, 'RelatedPerson'
+
+      # @!parse RESEARCHER = Researcher
+      new :RESEARCHER, 'Researcher'
+
+      # @!parse RESEARCH_GROUP = ResearchGroup
+      new :RESEARCH_GROUP, 'ResearchGroup'
+
+      # @!parse RIGHTS_HOLDER = RightsHolder
+      new :RIGHTS_HOLDER, 'RightsHolder'
+
+      # @!parse SPONSOR = Sponsor
+      new :SPONSOR, 'Sponsor'
+
+      # @!parse SUPERVISOR = Supervisor
+      new :SUPERVISOR, 'Supervisor'
+
+      # @!parse WORK_PACKAGE_LEADER = WorkPackageLeader
+      new :WORK_PACKAGE_LEADER, 'WorkPackageLeader'
+
+      # @!parse OTHER = Other
+      new :OTHER, 'Other'
+
+    end
+
+    # The institution or person responsible for collecting, creating, or otherwise contributing to the developement of the dataset.
+    class Contributor
+      include XML::Mapping
+
+      # @!attribute [rw] name
+      #   @return [String] the personal name of the contributor, in the format `Family, Given`. Cannot be empty or nil
+      text_node :name, 'contributorName'
+
+      # @!attribute [rw] identifier
+      #   @return [NameIdentifier, nil] an identifier for the contributor. Optional.
+      object_node :identifier, 'nameIdentifier', class: NameIdentifier, default_value: nil
+
+      # @!attribute [rw] affiliations
+      #   @return [Array<String>] the contributor's affiliations. Defaults to an empty list.
+      array_node :affiliations, 'affiliation', class: String, default_value: []
+
+      # @!attribute [rw] type
+      #   @return [ContributorType] the contributor type. Cannot be nil.
+      typesafe_enum_node :type, '@contributorType', class: ContributorType
+
+      alias_method :_name=, :name=
+      alias_method :_type=, :type=
+      private :_name=
+      private :_type=
+
+      # Initializes a new {Contributor}.
+      # @param name [String] the personal name of the contributor, in the format `Family, Given`. Cannot be empty or nil
+      # @param identifier [NameIdentifier, nil] an identifier for the contributor. Optional.
+      # @param affiliations [Array<Affiliation>] the contributor's affiliations. Defaults to an empty list.
+      # @param type [ContributorType] the contributor type. Cannot be nil.
+      def initialize(name:, identifier: nil, affiliations: nil, type:)
+        self.name = name
+        self.identifier = identifier
+        self.affiliations = affiliations || []
+        self.type = type
+      end
+
+      def name=(value)
+        fail ArgumentError, 'Name cannot be empty or nil' unless value && !value.empty?
+        self._name = value
+      end
+
+      def type=(value)
+        fail ArgumentError, 'Type cannot be nil' unless value
+        self._type = value
+      end
+    end
+  end
+end

data/lib/datacite/mapping/creator.rb

@@ -0,0 +1,48 @@
+require 'xml/mapping'
+require_relative 'name_identifier'
+
+module Datacite
+  module Mapping
+    # The main researchers involved working on the data, or the authors of the publication in priority order.
+    class Creator
+      include XML::Mapping
+
+      # @!attribute [rw] name
+      #   @return [String] The personal name of the creator, in the format `Family, Given`. Cannot be empty or nil.
+      text_node :name, 'creatorName'
+
+      # @!attribute [rw] identifier
+      #   @return [NameIdentifier, nil] An identifier for the creator. Optional.
+      object_node :identifier, 'nameIdentifier', class: NameIdentifier, default_value: nil
+
+      # @!attribute [rw] affiliations
+      #   @return [Array<Affiliation>, nil] The creator's affiliations. Defaults to an empty list.
+      array_node :affiliations, 'affiliation', class: String, default_value: []
+
+      alias_method :_name=, :name=
+      private :_name=
+
+      alias_method :_affiliations=, :affiliations=
+      private :_affiliations=
+
+      # Initializes a new {Creator}.
+      # @param name [String] The personal name of the creator, in the format `Family, Given`. Cannot be empty or nil.
+      # @param identifier [NameIdentifier, nil] An identifier for the creator. Optional.
+      # @param affiliations [Array<Affiliation>, nil] The creator's affiliations. Defaults to an empty list.
+      def initialize(name:, identifier: nil, affiliations: [])
+        self.name = name
+        self.identifier = identifier
+        self.affiliations = affiliations
+      end
+
+      def name=(value)
+        fail ArgumentError, 'Name cannot be empty or nil' unless value && !value.empty?
+        self._name = value
+      end
+
+      def affiliations=(value)
+        self._affiliations = value || []
+      end
+    end
+  end
+end

data/lib/datacite/mapping/date.rb

@@ -0,0 +1,153 @@
+require 'xml/mapping_extensions'
+
+module Datacite
+  module Mapping
+
+    # Controlled vocabulary of date types.
+    class DateType < TypesafeEnum::Base
+      # @!parse ACCEPTED = Accepted
+      new :ACCEPTED, 'Accepted'
+
+      # @!parse AVAILABLE = Available
+      new :AVAILABLE, 'Available'
+
+      # @!parse COPYRIGHTED = Copyrighted
+      new :COPYRIGHTED, 'Copyrighted'
+
+      # @!parse COLLECTED = Collected
+      new :COLLECTED, 'Collected'
+
+      # @!parse CREATED = Created
+      new :CREATED, 'Created'
+
+      # @!parse ISSUED = Issued
+      new :ISSUED, 'Issued'
+
+      # @!parse SUBMITTED = Submitted
+      new :SUBMITTED, 'Submitted'
+
+      # @!parse UPDATED = Updated
+      new :UPDATED, 'Updated'
+
+      # @!parse VALID = Valid
+      new :VALID, 'Valid'
+
+    end
+
+    # Represents a DataCite `<date/>` field, which can be a year, date (year-month-day or just year-month),
+    # or ISO8601 datetime.
+    #
+    # @!attribute [r] year
+    #   @return [Integer] The year.
+    # @!attribute [r] month
+    #   @return [Integer, nil] The month. Can be `nil` if no month was specified.
+    # @!attribute [r] day
+    #   @return [Integer, nil] The day. Can be `nil` if no day was specified.
+    # @!attribute [r] hour
+    #   @return [Integer, nil] The hour. Can be `nil` if no hour was specified.
+    # @!attribute [r] minute
+    #   @return [Integer, nil] The minutes. Can be `nil` if no minutes were specified.
+    # @!attribute [r] sec
+    #   @return [Integer, nil] The seconds. Can be `nil` if no seconds were specified.
+    # @!attribute [r] nsec
+    #   @return [Integer, nil] The nanoseconds. Can be `nil` if no nanoseconds were specified.
+    class Date
+      include XML::Mapping
+
+      # @!attribute [rw] type
+      #  @return [DateType] the type of date. Cannot be nil.
+      typesafe_enum_node :type, '@dateType', class: DateType
+      alias_method :_type=, :type=
+      private :_type=
+
+      # @!method value
+      #   @return [String] The value as a string. May be any [W3C DateTime format](http://www.w3.org/TR/NOTE-datetime).
+      text_node :value, 'text()'
+      alias_method :_value=, :value=
+      private :_value=
+
+      attr_reader :year
+      attr_reader :month
+      attr_reader :day
+      attr_reader :hour
+      attr_reader :minute
+      attr_reader :sec
+      attr_reader :nsec
+
+      # Initializes a new `Date`
+      #
+      # @param type [DateType] the type of date. Cannot be nil.
+      # @param value [DateTime, Date, Integer, String] The value, as a `DateTime`, `Date`, or `Integer`,
+      #   or as a `String` in any [W3C DateTime format](http://www.w3.org/TR/NOTE-datetime)
+      def initialize(type:, value:)
+        self.type = type
+        self.value = value
+      end
+
+      def type=(val)
+        fail ArgumentError, 'Date type cannot be nil' unless val
+        self._type = val
+      end
+
+      # Sets the value.
+      #
+      # @param val [DateTime, Date, Integer, String] The value, as a `DateTime`, `Date`, or `Integer`,
+      #   or as a `String` in any [W3C DateTime format](http://www.w3.org/TR/NOTE-datetime)
+      def value=(val) # rubocop:disable Metrics/MethodLength, Metrics/AbcSize, Metrics/CyclomaticComplexity, Metrics/PerceivedComplexity
+        @date_time = to_datetime(val)
+        @date = @date_time ? @date_time.to_date : to_date(val)
+        @year = to_year(val)
+        @month = to_month(val)
+        @day = to_day(val)
+        new_value = val.respond_to?(:iso8601) ? val.iso8601 : val.to_s
+        if new_value.include?('T')
+          @hour = @date_time.hour if @date_time
+          @minute = @date_time.minute if @date_time
+          @sec = @date_time.sec if @date_time
+          @nsec = @date_time.to_time.nsec if @date_time
+        end
+        fail ArgumentError, "Unable to parse date value '#{val}'" unless @year
+        self._value = new_value
+      end
+
+      private
+
+      def to_year(val)
+        return val if val.is_a?(Integer)
+        return val.year if val.respond_to?(:year)
+        matchdata = val.to_s.match(/^[0-9]+/)
+        matchdata[0].to_i if matchdata
+      end
+
+      def to_month(val)
+        return val.month if val.respond_to?(:month)
+        matchdata = val.to_s.match(/^[0-9]+-([0-9]+)/)
+        matchdata[1].to_i if matchdata
+      end
+
+      def to_day(val)
+        return val.day if val.respond_to?(:day)
+        matchdata = val.to_s.match(/^[0-9]+-[0-9]+-([0-9]+)/)
+        matchdata[1].to_i if matchdata
+      end
+
+      def to_datetime(val)
+        return val if val.is_a?(DateTime)
+        DateTime.parse(val.to_s)
+      rescue ArgumentError => e
+        Mapping.log.debug("Can't extract DateTime from date value '#{val}': #{e}")
+        nil
+      end
+
+      def to_date(val)
+        return val if val.is_a?(::Date)
+        return ::Date.parse(val.iso8601) if val.respond_to?(:iso8601)
+        ::Date.parse(val.to_s)
+      rescue ArgumentError => e
+        Mapping.log.debug("Can't extract Date from date value '#{val}': #{e}")
+        nil
+      end
+
+    end
+  end
+end