marc2linkeddata 0.0.1

data/lib/marc2linkeddata/parseMarcAuthority.rb

@@ -0,0 +1,492 @@
+
+# Marc21 Authority fields are documented at
+# http://www.loc.gov/marc/authority/ecadlist.html
+# http://www.loc.gov/marc/authority/ecadhome.html
+
+module Marc2LinkedData
+
+  class ParseMarcAuthority
+
+    # TODO: provide iterator pattern on an entire file of records.
+    # @leader = ParseMarcAuthority::parse_leader(marc_file)
+    # raw = marc_file.read(@leader[:length])
+    # @record = MARC::Reader.decode(raw)
+
+    @@config = nil
+
+    attr_reader :loc
+    attr_reader :isni
+    attr_reader :viaf
+
+    def initialize(record)
+      @@config ||= Marc2LinkedData.configuration
+      @record = record
+      @graph = RDF::Graph.new
+      @loc = nil
+      @isni = nil
+      @viaf = nil
+    end
+
+    def get_fields(field_num)
+      fields = @record.fields.select {|f| f if f.tag == field_num }
+      raise "Invalid data in field #{field_num}" if fields.length < 1
+      fields
+    end
+
+    # Try to use the SUL catkey and/or the OCLC control numbers, maybe SUL
+    # catkey in the record IRI
+    def get_id
+      # extract ID from control numbers, see
+      # http://www.loc.gov/marc/authority/ad001.html
+      #field001 = record.fields.select {|f| f if f.tag == '001' }.first.value
+      #field003 = record.fields.select {|f| f if f.tag == '003' }.first.value
+      #"#{field003}-#{field001}"
+      get_fields('001').first.value
+    end
+
+    def get_iri(field, iri_pattern)
+      begin
+        iris = field.subfields.collect {|f| f.value if f.value.include? iri_pattern }
+        iris.first || nil
+      rescue
+        nil
+      end
+    end
+
+    def get_iri4isni
+      isni_iri = nil
+      begin
+        # e.g. http://www.isni.org/0000000109311081
+        field = get_fields(@@config.field_auth_isni).first
+        isni_iri = get_iri(field, 'isni.org')
+        # If ISNI is not already in the MARC record, try to get it from VIAF.
+        if isni_iri.nil? && @@config.get_isni
+          isni_iri = @viaf.get_isni rescue nil
+          @@config.logger.debug 'Failed to resolve ISNI URI' if isni_iri.nil?
+          # binding.pry if @viaf.iri.to_s.include? '67737121' #@@config.debug
+        end
+        unless isni_iri.nil?
+          # Ensure the ISNI IRI has this prefix: http://www.isni.org/isni/
+          isni_iri.gsub('www.isni.org', 'www.isni.org/isni') unless isni_iri =~ /www\.isni\.org\/isni\//
+        end
+        return isni_iri
+      rescue
+        nil
+      end
+    end
+
+    def get_iri4lib
+      "#{@@config.prefixes['lib_auth']}#{get_id}"
+    end
+
+    def get_iri4loc
+      loc_iri = nil
+      begin
+        # e.g. http://id.loc.gov/authorities/names/n42000906
+        field = get_fields(@@config.field_auth_loc).first
+        loc_iri = get_iri(field, 'id.loc.gov')
+      rescue
+      end
+      begin
+        if loc_iri.nil?
+          # If the LOC is not in the marc record, try to determine the LOC IRI from the ID.
+          loc_id = get_id
+          if loc_id =~ /^n/i
+            loc_iri = "#{@@config.prefixes['loc_names']}#{loc_id.downcase}"
+          end
+          if loc_id =~ /^sh/i
+            loc_iri = "#{@@config.prefixes['loc_subjects']}#{loc_id.downcase}"
+          end
+          unless loc_iri.nil?
+            # Verify the URL (used HEAD so it's as fast as possible)
+            @@config.logger.debug "Trying to validate LOC IRI: #{loc_iri}"
+            res = Marc2LinkedData.http_head_request(loc_iri + '.rdf')
+            case res.code
+              when '200'
+                # it's good to go
+              when '301'
+                # use the redirection
+                loc_iri = res['location']
+              when '302','303'
+                #302 Moved Temporarily
+                #303 See Other
+                # Use the current URL, most get requests will follow a 302 or 303
+              else
+                loc_iri = nil
+            end
+          end
+          if loc_iri.nil?
+            # If it gets here, it's a problem.
+            binding.pry if @@config.debug
+            @@config.logger.error 'FAILURE to resolve LOC IRI'
+          else
+            @@config.logger.debug "DISCOVERED LOC IRI: #{loc_iri}"
+          end
+        else
+          @@config.logger.debug "MARC contains LOC IRI: #{loc_iri}"
+        end
+        return loc_iri
+      rescue
+        nil
+      end
+    end
+
+    def get_iri4oclc
+      begin
+        field = get_fields(@@config.field_auth_oclc).first
+        oclc_cn = field.subfields.collect {|f| f.value if f.code == 'a'}.first
+        oclc_id = /\d+$/.match(oclc_cn).to_s
+        oclc_id.empty? ? nil : "http://www.worldcat.org/oclc/#{oclc_id}"
+      rescue
+        nil
+      end
+    end
+
+    def get_iri4viaf
+      begin
+        # e.g. http://viaf.org/viaf/181829329
+        # VIAF RSS feed for changes, e.g. http://viaf.org/viaf/181829329.rss
+        field = get_fields(@@config.field_auth_viaf).first
+        viaf_iri = get_iri(field, 'viaf.org')
+        # If VIAF is not already in the MARC record, try to get from LOC.
+        if viaf_iri.nil? && @@config.get_viaf
+          viaf_iri = @loc.get_viaf rescue nil
+          @@config.logger.debug 'Failed to resolve VIAF URI' if viaf_iri.nil?
+        end
+        return viaf_iri
+      rescue
+        nil
+      end
+    end
+
+    def self.parse_leader(file_handle, leader_bytes=24)
+      # example:
+      #record.leader
+      #=> "00774cz  a2200253n  4500"
+      # 00-04: '00774' - record length
+      # 05:    'c' - corrected or revised
+      # 06:    'z' - always 'z' for authority records
+      # 09:    'a' - UCS/Unicode
+      # 12-16: '00253' - base address of data, Length of Leader and Directory
+      # 17:    'n' - Complete authority record
+      # leader_status_codes = {
+      #     'a' => 'Increase in encoding level',
+      #     'c' => 'Corrected or revised',
+      #     'd' => 'Deleted',
+      #     'n' => 'New',
+      #     'o' => 'Obsolete',
+      #     's' => 'Deleted; heading split into two or more headings',
+      #     'x' => 'Deleted; heading replaced by another heading'
+      # }
+      leader = file_handle.read(leader_bytes)
+      file_handle.seek(-1 * leader_bytes, IO::SEEK_CUR)
+      {
+          :length => leader[0..4].to_i,
+          :status => leader[5],  # leader_status_codes[ record.leader[5] ]
+          :type => leader[6],    # always 'z' for authority records
+          :encoding => leader[9],  # translate letter code into ruby encoding string
+          :data_address => leader[12..16].to_i,
+          :complete => leader[17].include?('n')
+      }
+    end
+
+    def parse_008
+      # http://www.loc.gov/marc/authority/concise/ad008.html
+      field = get_fields('008').first
+      field008 = field.value
+      languages = []
+      languages.append('English') if ['b','e'].include? field008[8]
+      languages.append('French') if ['b','f'].include? field008[8]
+      rules = ''
+      rules = 'EARLIER' if field008[10] == 'a'
+      rules = 'AACR1' if field008[10] == 'b'
+      rules = 'AACR2' if field008[10] == 'c'
+      rules = 'AACR2 compatible' if field008[10] == 'd'
+      rules = 'OTHER' if field008[10] == 'z'
+      rules = 'N/A' if field008[10] == 'n'
+      # 32 - Undifferentiated personal name
+      # Whether the personal name in a name or name/title heading contained in field 100 in an established heading record or a reference record is used by one person or by two or more persons.
+      # a - Differentiated personal name
+      #     Personal name in field 100 is a unique name.
+      # b - Undifferentiated personal name
+      #     Personal name in field 100 is used by two or more persons.
+      # n - Not applicable
+      #     1XX heading is not a personal name or the personal name is a family name.
+      # | - No attempt to code
+      {
+          :date => Date.strptime(field008[0..5], "%y%m%d"),
+          :geographic_subdivision => field008[6], # '#', d, i, n, or '|'
+          :romanization_scheme => field008[7], # a..g, n, or '|'
+          :languages => languages,
+          :kind => field008[9], # a..g, or '|'
+          :rules => rules,
+          :heading_system => field008[11],
+          :series_type => field008[12],
+          :series_numbered => field008[13],
+          :use_1XX_for_7XX => field008[14] == 'a',
+          :use_1XX_for_6XX => field008[15] == 'a',
+          :use_1XX_for_4XX => field008[16] == 'a',
+          :use_1XX_for_8XX => field008[16] == 'a',
+          :type_subject_subdivision => field008[17],
+          # 18-27 - Undefined character positions
+          :type_government_agency => field008[28],
+          :reference_evaluation => field008[29],
+          # 30 - Undefined character position
+          :record_available => field008[31] == 'a',
+          # TODO: 32
+          # TODO: 33
+          # 34-37 - Undefined character positions
+          # TODO: 38
+          # TODO: 39
+      }
+    end
+
+    def parse_100
+      # http://www.loc.gov/marc/authority/concise/ad100.html
+      begin
+        # 100 is a personal name
+        field = get_fields('100').first
+        # field = @record.fields.select {|f| f if f.tag == '100' }.first
+        name = field.subfields.select {|f| f.code == 'a' }.first.value rescue ''
+        name.force_encoding('UTF-8')
+      rescue
+        'ERROR_PERSONAL_NAME'
+      end
+    end
+
+    def parse_110
+      # http://www.loc.gov/marc/authority/concise/ad110.html
+      begin
+        # 110 is a corporate name
+        field = @record.fields.select {|f| f if f.tag == '110' }.first
+        a = field.subfields.collect {|f| f.value if f.code == 'a' }.compact rescue []
+        b = field.subfields.collect {|f| f.value if f.code == 'b' }.compact rescue []
+        c = field.subfields.collect {|f| f.value if f.code == 'c' }.compact rescue []
+        name = [a,b,c].flatten.join(' : ')
+        name.force_encoding('UTF-8')
+      rescue
+        'ERROR_CORPORATE_NAME'
+      end
+    end
+
+    def parse_111
+      # http://www.loc.gov/marc/authority/concise/ad111.html
+      begin
+        # 111 is a meeting name
+        field = @record.fields.select {|f| f if f.tag == '111' }.first
+        a = field.subfields.collect {|f| f.value if f.code == 'a' }.compact rescue []
+        # TODO: incorporate additional subfields?
+        # b = field.subfields.collect {|f| f.value if f.code == 'b' }.compact rescue []
+        # c = field.subfields.collect {|f| f.value if f.code == 'c' }.compact rescue []
+        # name = [a,b,c].flatten.join(' : ')
+        # name.force_encoding('UTF-8')
+        a.force_encoding('UTF-8')
+      rescue
+        'ERROR_MEETING_NAME'
+      end
+    end
+
+    def parse_151
+      # http://www.loc.gov/marc/authority/concise/ad151.html
+      begin
+        # 151 is a geographic name
+        field = @record.fields.select {|f| f if f.tag == '151' }.first
+        name = field.subfields.collect {|f| f.value if f.code == 'a' }.first rescue ''
+        name.force_encoding('UTF-8')
+      rescue
+        'ERROR_PLACE_NAME'
+      end
+    end
+
+    def parse_loc_auth_name
+      #
+      # Create triples for various kinds of LOC authority.
+      # At present, this relies on LOC RDF to differentiate
+      # types of authorities.  It should be possible to do this
+      # from the MARC directly, if @@config.get_loc is false.
+      #
+      # The MARC data differentiates them according to the tag number.
+      # The term 'name' refers to:
+      #  X00 - Personal Name
+      #  X10 - Corporate Name
+      #  X11 - Meeting Name
+      #  X30 - Uniform Title
+      #  X51 - Jurisdiction / Geographic Name
+      #
+      @@config.logger.warn "LOC URL: #{@loc.iri} DEPRECATED" if @loc.deprecated?
+      name = ''
+      if @loc.conference?
+        # e.g. http://id.loc.gov/authorities/names/n79044866
+        name = @loc.label || parse_111
+        @graph.insert RDF::Statement(@lib.rdf_uri, RDF.type, RDF::SCHEMA.event)
+      elsif @loc.corporation?
+        name = @loc.label || parse_110
+        @graph.insert RDF::Statement(@lib.rdf_uri, RDF.type, RDF::FOAF.Organization) if @@config.use_foaf
+        @graph.insert RDF::Statement(@lib.rdf_uri, RDF.type, RDF::SCHEMA.Organization) if @@config.use_schema
+      elsif @loc.name_title?
+        # e.g. http://id.loc.gov/authorities/names/n79044934
+        # Skipping these, because the person entity should be in
+        # an additional record and we don't want the title content.
+        binding.pry if @@config.debug
+        return ''
+      elsif @loc.person?
+        name = @loc.label || parse_100
+        @graph.insert RDF::Statement(@lib.rdf_uri, RDF.type, RDF::FOAF.Person) if @@config.use_foaf
+        @graph.insert RDF::Statement(@lib.rdf_uri, RDF.type, RDF::SCHEMA.Person) if @@config.use_schema
+        # VIAF extracts first and last name, try to use them. Note
+        # that VIAF uses schema:name, schema:givenName, and schema:familyName.
+        if @@config.get_viaf && ! @viaf.nil?
+          @viaf.family_names.each do |n|
+            # ln = URI.encode(n)
+            # TODO: try to get a language type, if VIAF provide it.
+            # name = RDF::Literal.new(n, :language => :en)
+            ln = RDF::Literal.new(n)
+            @graph.insert RDF::Statement(@lib.rdf_uri, RDF::FOAF.familyName, ln) if @@config.use_foaf
+            @graph.insert RDF::Statement(@lib.rdf_uri, RDF::SCHEMA.familyName, ln) if @@config.use_schema
+          end
+          @viaf.given_names.each do |n|
+            # fn = URI.encode(n)
+            # TODO: try to get a language type, if VIAF provide it.
+            # name = RDF::Literal.new(n, :language => :en)
+            fn = RDF::Literal.new(n)
+            @graph.insert RDF::Statement(@lib.rdf_uri, RDF::FOAF.firstName, fn) if @@config.use_foaf
+            @graph.insert RDF::Statement(@lib.rdf_uri, RDF::SCHEMA.givenName, fn) if @@config.use_schema
+          end
+        end
+      elsif @loc.place?
+        # e.g. http://id.loc.gov/authorities/names/n79045127
+        name = @loc.label || parse_151
+        @graph.insert RDF::Statement(@lib.rdf_uri, RDF.type, RDF::SCHEMA.Place)
+      else
+        # TODO: find out what type this is.
+        binding.pry if @@config.debug
+        name = @loc.label || ''
+        # Note: schema.org has no immediate parent for Person or Organization
+        @graph.insert RDF::Statement(@lib.rdf_uri, RDF.type, RDF::FOAF.Agent) if @@config.use_foaf
+        @graph.insert RDF::Statement(@lib.rdf_uri, RDF.type, RDF::SCHEMA.Thing) if @@config.use_schema
+      end
+      if name != ''
+        # name_encoding = URI.encode(name)
+        name = RDF::Literal.new(name)
+        @graph.insert RDF::Statement(@lib.rdf_uri, RDF::FOAF.name, name) if @@config.use_foaf
+        @graph.insert RDF::Statement(@lib.rdf_uri, RDF::SCHEMA.name, name) if @@config.use_schema
+      end
+    end
+
+    def parse_loc_auth_subject
+      # TODO: what to do with subjects?
+      binding.pry if @@config.debug
+      # The term 'subject' refers to:
+      #  X30 - Uniform Titles
+      #  X48 - Chronological Terms
+      #  X50 - Topical Terms
+      #  X51 - Geographic Names
+      #  X55 - Genre/Form Terms
+      #
+      # The term 'subject subdivision' refers to:
+      # X80 - general subdivision terms
+      # X81 - geographic subdivision names
+      # X82 - chronological subdivision terms
+      # X85 - form subdivision terms
+    end
+
+    def get_oclc_links
+      oclc_iri = nil
+      begin
+        # Try to get OCLC using LOC ID.
+        oclc_iri = @loc.get_oclc_identity
+      rescue
+        # Try to get OCLC using 035a field data, but
+        # this is not as reliable/accurate as LOC.
+        oclc_iri = get_iri4oclc
+      end
+      unless oclc_iri.nil?
+        # Try to get additional data from OCLC, using the RDFa
+        # available in the OCLC identities pages.
+        oclc_auth = OclcIdentity.new oclc_iri
+        @graph.insert RDF::Statement(@loc.rdf_uri, RDF::OWL.sameAs, oclc_auth.rdf_uri)
+        oclc_auth.creative_works.each do |creative_work_uri|
+          # Notes on work-around for OCLC data inconsistency:
+          # RDFa for http://www.worldcat.org/identities/lccn-n79044798 contains:
+          # <http://worldcat.org/oclc/747413718> a <http://schema.org/CreativeWork> .
+          # However, the RDF for <http://worldcat.org/oclc/747413718> contains:
+          # <http://www.worldcat.org/oclc/747413718> schema:exampleOfWork <http://worldcat.org/entity/work/id/994448191> .
+          # Note how the subject here is 'WWW.worldcat.org' instead of 'worldcat.org'.
+          #creative_work_iri = creative_work.to_s.gsub('worldcat.org','www.worldcat.org')
+          #creative_work_iri = creative_work_iri.gsub('wwwwww','www') # in case it gets added already by OCLC
+          creative_work = OclcCreativeWork.new creative_work_uri
+          @graph.insert RDF::Statement(oclc_auth.rdf_uri, RDF::RDFS.seeAlso, creative_work.rdf_uri)
+          if @@config.oclc_auth2works
+            # Try to use VIAF to relate auth to work as creator, contributor, editor, etc.
+            # Note that this requires additional RDF retrieval for each work (slower processing).
+            unless @viaf.nil?
+              if creative_work.creator? @viaf.iri
+                @graph.insert RDF::Statement(creative_work.rdf_uri, RDF::SCHEMA.creator, oclc_auth.rdf_uri)
+              elsif creative_work.contributor? @viaf.iri
+                @graph.insert RDF::Statement(creative_work.rdf_uri, RDF::SCHEMA.contributor, oclc_auth.rdf_uri)
+              elsif creative_work.editor? @viaf.iri
+                @graph.insert RDF::Statement(creative_work.rdf_uri, RDF::SCHEMA.editor, oclc_auth.rdf_uri)
+              end
+            end
+            # TODO: Is auth the subject of the work (as in biography) or both (as in autobiography).
+            # binding.pry if @@config.debug
+            # binding.pry if creative_work.iri.to_s == 'http://www.worldcat.org/oclc/006626542'
+            # Try to find the generic work entity for this example work.
+            creative_work.get_works.each do |oclc_work_uri|
+              oclc_work = OclcWork.new oclc_work_uri
+              @graph.insert RDF::Statement(creative_work.rdf_uri, RDF::SCHEMA.exampleOfWork, oclc_work.rdf_uri)
+            end
+          end
+        end
+      end
+    end
+
+    # TODO: use an 'affiliation' entry, maybe 373?  (optional field)
+
+    def to_ttl
+      graph.to_ttl
+    end
+
+    def graph
+      # TODO: figure out how to specify all the graph prefixes.
+      return @graph unless @graph.empty?
+      @lib = LibAuth.new get_iri4lib
+      # Try to find LOC, VIAF, and ISNI IRIs in the MARC record
+      @loc = Loc.new get_iri4loc rescue nil
+      # Try to identify problems in getting an LOC IRI.
+      binding.pry if (@@config.debug && @loc.nil?)
+      # might require LOC to get ISNI.
+      @viaf = Viaf.new get_iri4viaf rescue nil
+      # might require VIAF to get ISNI.
+      @isni = Isni.new get_iri4isni rescue nil
+
+      # TODO: ORCID? VIVO? VITRO? Stanford CAP?
+
+      # Get LOC control number and add catalog permalink? e.g.
+      # http://lccn.loc.gov/n79046291
+
+      @graph.insert RDF::Statement(@lib.rdf_uri, RDF::OWL.sameAs, @loc.rdf_uri)
+      @graph.insert RDF::Statement(@lib.rdf_uri, RDF::OWL.sameAs, @viaf.rdf_uri) unless @viaf.nil?
+      @graph.insert RDF::Statement(@lib.rdf_uri, RDF::OWL.sameAs, @isni.rdf_uri) unless @isni.nil?
+      return @graph unless @@config.get_loc
+
+      # TODO: find codes in the marc record to differentiate the authority into
+      # TODO: person, organization, event, etc. without getting LOC RDF.
+
+      if @loc.iri.to_s =~ /name/
+        parse_loc_auth_name
+      elsif @loc.iri.to_s =~ /subjects/
+        parse_loc_auth_subject
+      else
+        binding.pry if @@config.debug
+      end
+      # Optional elaboration of authority data with OCLC identity and works.
+      get_oclc_links if @@config.get_oclc
+
+      @@config.logger.info "Extracted #{@loc.id}"
+      @graph
+    end
+  end
+
+end
+

data/lib/marc2linkeddata/readMarcCatalog.rb

@@ -0,0 +1,175 @@
+#!/usr/bin/env ruby
+
+require 'marc'
+require 'linkeddata'
+require 'pry'
+
+#EXAMPLE_RECORD_FILE='../marc/catalog/stf.00.mrc'
+EXAMPLE_RECORD_FILE='../marc/catalog/stf.51.mrc'
+
+## reading records from a batch file
+#reader = MARC::Reader.new(EXAMPLE_RECORD_FILE, :external_encoding => "MARC-8")
+#reader = MARC::Reader.new(EXAMPLE_RECORD_FILE, :external_encoding => "UTF-8", :validate_encoding => true)
+
+#reader = MARC::ForgivingReader.new(EXAMPLE_RECORD_FILE)
+
+handle = File.new(EXAMPLE_RECORD_FILE)
+#=> #<File:marc/authority/stf_auth.00.mrc>
+rec_length = handle.read(5).to_i
+#=> 774
+handle.rewind
+raw = handle.read(rec_length)
+record = MARC::Reader.decode(raw)
+
+# From http://www.loc.gov/marc/authority/adleader.html
+# System-Generated Elements - The following Leader elements are usually system generated:
+#
+# 00-04 	Logical record length
+#
+#    05 - Record status:
+#     a - Increase in encoding level
+#     c - Corrected or revised
+#     d - Deleted
+#     n - New
+#     o - Obsolete
+#     s - Deleted; heading split into two or more headings
+#     x - Deleted; heading replaced by another heading
+#
+#    06 - Type of record
+#     z - Authority data
+#
+# 07-08 	Undefined character positions
+#
+#    09 - Character coding scheme
+#     # - MARC-8
+#     a - UCS/Unicode
+#
+#    10 	Indicator count
+#     2 - Number of character positions used for indicators
+#
+#    11 	Subfield code count
+#     2 - Number of character positions used for a subfield code
+#
+# 12-16 	Base address of data
+#     [number] - Length of Leader and Directory
+#
+#    17 - Encoding level
+#     n - Complete authority record
+#     o - Incomplete authority record
+#
+# 20-23 	Entry map
+#
+# 18-19 - Undefined character positions
+#
+#    20 - Length of the length-of-field portion
+#     4 - Number of characters in the length-of-field portion of a Directory entry
+#
+#    21 - Length of the starting-character-position portion
+#     5 - Number of characters in the starting-character-position portion of a Directory entry
+#
+#    22 - Length of the implementation-defined portion
+#     0 - Number of characters in the implementation-defined portion of a Directory entry
+#
+# It is common for default values in other Leader elements to be generated automatically as well.
+# Capitalization - Alphabetic codes are input as lower case letters.
+#
+# example:
+#record.leader
+#=> "00774cz  a2200253n  4500"
+# 00-04: '00774' - record length
+# 05:    'c' - corrected or revised
+# 06:    'z' - always 'z' for authority records
+# 09:    'a' - UCS/Unicode
+# 12-16: '00253' - base address of data, Length of Leader and Directory
+# 17:    'n' - Complete authority record
+LEADER_STATUS_CODES = {
+  'a' => 'Increase in encoding level',
+  'c' => 'Corrected or revised',
+  'd' => 'Deleted',
+  'n' => 'New',
+  'o' => 'Obsolete',
+  's' => 'Deleted; heading split into two or more headings',
+  'x' => 'Deleted; heading replaced by another heading'
+}
+def leader_parse(record)
+  leader = {
+    :length => record.leader[0..4].to_i,
+    :status => record.leader[5],  # LEADER_STATUS_CODES[ record.leader[5] ]
+    :type => record.leader[6],    # always 'z' for authority records
+    :encoding => record.leader[9],  # TODO: translate letter code into ruby encoding string
+    :data_address => record.leader[12..16].to_i,
+    :complete => record.leader[17].include?('n')
+  }
+end
+
+
+# Stanford Resource keys and Ckeys can collide. They are only unique within their own set.
+#
+# When I do a catalogdump for ckey 6809804 I see:
+#
+# .948.   |hNO HOLDINGS IN STF - 7 OTHER HOLDINGS
+#
+# When we do a catalogdump for searchworks we filter the results to only export
+# records with holdings, and not those things which are on order or "shadowed"
+# i.e. hidden from public view, although we still have the bibliographic data in
+# the database. When I extracted the records for conversion I selected all of
+# them.
+#
+# - Josh
+
+
+# Create SUL LOD...
+SUL_URI = RDF::URI.new('http://linked-data.stanford.edu/library/')
+
+# extract catalog key from field 001 (use the first one)
+field001 = record.fields.select {|f| f if f.tag == '001' }.first
+cat_key = field001.value.strip
+CAT_URI = SUL_URI.join("catalog/#{cat_key}")
+
+# TODO: Evaluate whether cat_key is in SearchWorks, e.g.
+# "http://searchworks.stanford.edu/catalog/#{cat_key}"
+# http://searchworks.stanford.edu/catalog/7106054
+
+# TODO: extract 035a for OCLC master control number.
+# TODO: map the OCLC to the OCLC work number.
+field035 = record.fields.select {|f| f if f.tag == '035' }
+
+
+binding.pry
+exit!
+
+
+#There is nothing in the MARC record itself to indicate that a holding is
+#'shadowed' (not available for public view), but one idea to handle them is
+#to supply a list of shadowed ckeys and that list could easily be transformed
+#
+#into a list of triples like this:
+#	<http://linked-data.stanford.edu/library/catalog/{cKey}> <rdf:Property> <http://linked-data.stanford.edu/library/catalog/isShadowed>
+#...or what ever predicate and object you want to use. Then you can load those into the triple store.
+
+# .948.   |hNO HOLDINGS IN STF - 7 OTHER HOLDINGS
+# When we do a catalogdump for searchworks we filter the results to only export
+# records with holdings, and not those things which are on order or "shadowed"
+# i.e. hidden from public view, although we still have the bibliographic data in
+# the database. When I extracted the records for conversion I selected all of
+# them.
+field948 = record.fields.select {|f| f if f.tag == '948' }
+holdings = field948.first.value
+
+
+# TODO: construct RDF model, see http://blog.datagraph.org/2010/03/rdf-for-ruby
+# RDF::Literal.new("Hello!", :language => :en)
+#
+lod = {
+  :id => cat_uri,
+  :oclc => oclc4loc.collect {|uri| RDF::URI.new(uri) },
+}
+
+binding.pry
+exit!
+
+#for record in reader
+#  # print out field 245 subfield a
+#  puts record['245']['a']
+#end
+