marc4r 0.0.1

checksums.yaml

@@ -0,0 +1,7 @@
+---
+SHA1:
+  metadata.gz: d0a58b8d188bfed7b29da6dd0a2924a2364e0b98
+  data.tar.gz: f61c55be658daf6e9e74e49f8b52b9a2792a6334
+SHA512:
+  metadata.gz: 6033e3fe5eb6d220953208ad7252c81f24fb757af0dc8c965a02ec0bba12da9b9e24286d497726a0bd802c2381416222fa3b79944446235e629c87d3cf9bd2a6
+  data.tar.gz: e2693eca537b27f9c69d5cbfe7a6744a765a0e5d8bd34e14e50d9e1be283a444374600639f0bc703e0243acf8dfcebabafe79750bb5cfd13c32042b68769cb93

data/.gitignore

@@ -0,0 +1 @@
+*.java

data/Gemfile

@@ -0,0 +1,4 @@
+source 'https://rubygems.org'
+
+# Please see marc4r.gemspec for dependency information.
+gemspec

data/README.md

@@ -0,0 +1,4 @@
+Marc4R is a ruby MARC parser based on fork of Marc4J
+MARC Specification at http://www.loc.gov/marc/
+Marc4J Fork at https://github.com/marc4j/marc4j
+Marc4J Original Copyright 2005 Bas Peters, distributed under the LGPL

data/Rakefile

@@ -0,0 +1,13 @@
+#!/usr/bin/env rake
+
+begin
+  require 'bundler/setup'
+rescue LoadError
+  puts 'You must `gem install bundler` and `bundle install` to run rake tasks'
+end
+
+Bundler::GemHelper.install_tasks
+
+Dir.glob('tasks/*.rake').each { |r| import r }
+
+task default: :spec

data/VERSION

@@ -0,0 +1 @@
+0.0.1

data/lib/marc4r.rb

@@ -0,0 +1,24 @@
+# -*- encoding: utf-8 -*-
+module Marc4R
+  MARCXML_NS_URI = "http://www.loc.gov/MARC21/slim"
+  autoload :Fields, 'marc4r/fields'
+  autoload :Record, 'marc4r/record'
+  autoload :MarcStreamReader, 'marc4r/marc_stream_reader'
+  module Terminators
+  	RECORD_TERMINATOR = "\x1D"
+  	FIELD_TERMINATOR = "\x1E"
+  	SUBFIELD_TERMINATOR = "\x1F"
+  	BLANK = "\x20"
+  end
+  module Encodings
+    # MARC-8 ANSEL ENCODING
+    MARC_8_ENCODING = "MARC8"
+
+    # ISO5426 ENCODING
+    ISO5426_ENCODING = "ISO5426"
+
+    # ISO6937 ENCODING
+    ISO6937_ENCODING = "ISO6937"    
+  end
+  class MarcException < Exception; end
+end

data/lib/marc4r/fields.rb

@@ -0,0 +1,28 @@
+module Marc4R::Fields
+  autoload :ControlField, 'marc4r/fields/control_field'
+  autoload :DataField, 'marc4r/fields/data_field'
+  autoload :Subfield, 'marc4r/fields/subfield'
+  module Taggable
+    attr_accessor :tag
+    def <=>(compare)
+      if compare.is_a? Marc4R::Fields::Taggable
+        tag <=> compare.tag
+      else
+        super
+      end
+    end
+  end
+  module Datable
+    attr_accessor :data
+    def =~(pattern)
+      if data
+        data =~ pattern
+      else
+        nil
+      end
+    end
+  end
+  module Idable
+    attr_accessor :id
+  end
+end

data/lib/marc4r/fields/control_field.rb

@@ -0,0 +1,19 @@
+class Marc4R::Fields::ControlField
+  include Marc4R::Fields::Idable
+  include Marc4R::Fields::Taggable
+  include Marc4R::Fields::Datable
+  def initialize(tag=nil,data=nil)
+    self.tag= tag if tag
+    self.data= data if data
+  end
+  def self.accepts_tag(tag)
+    tag =~ /^00[0-9]$/
+  end
+  def eql? o
+  	if o.is_a? Marc4R::Fields::ControlField
+      (self.tag.eql? o.tag) && (self.data.eql? o.data)
+    else
+      false
+    end
+  end
+end

data/lib/marc4r/fields/data_field.rb

@@ -0,0 +1,46 @@
+class Marc4R::Fields::DataField
+  include Marc4R::Fields::Idable
+  include Marc4R::Fields::Taggable
+  attr_accessor :indicator1, :indicator2
+  def initialize(tag=nil, ind1=nil, ind2=nil, *args)
+    self.tag= tag if tag
+    self.indicator1 = ind1 if ind1
+    self.indicator2 = ind2 if ind2
+    @subfields = []
+    args.each_slice(2) do |slice|
+      if (slice[0].size != 1)
+        raise Marc4R::MarcException.new("subfield code must be a single character #{slice.inspect}")
+      end
+      @subfields << Marc4R::Fields::Subfield.new(*slice)
+    end
+  end
+
+  def subfields(code=nil)
+  	code ? @subfields.select {|sf| sf.code == code} : @subfields
+  end
+
+  def subfield(code=nil)
+  	subfields(code).first
+  end
+
+  def =~ pattern
+    subfields.detect {|f| f =~ pattern}
+  end
+
+  def eql? other
+    if other.is_a? Marc4R::Fields::DataField
+      r = self.tag.eql? other.tag
+      r &= self.indicator1.eql? other.indicator1
+      r &= self.indicator2.eql? other.indicator2
+      r &= self.subfields.size.eql? other.subfields.size
+      if r
+        self.subfields.each do |sf|
+          r &= other.subfields.detect {|osf| (osf.code.eql? sf.code) && (osf.data.eql? sf.data)}
+        end
+      end
+      r
+    else
+      false
+    end
+  end
+end

data/lib/marc4r/fields/subfield.rb

@@ -0,0 +1,12 @@
+class Marc4R::Fields::Subfield
+  include Marc4R::Fields::Idable
+  include Marc4R::Fields::Datable
+  attr_accessor :code
+  def initialize(code=nil, data=nil)
+    self.code = code if code
+    self.data = data if data
+  end
+  def to_s
+    "#{code}#{data}"
+  end
+end

data/lib/marc4r/marc_stream_reader.rb

@@ -0,0 +1,82 @@
+class Marc4R::MarcStreamReader
+  include Enumerable
+  def initialize(io, enc=nil)
+    @io = io
+    @encoding = enc ? Encoding.find(enc) : nil
+  end
+  def close
+    @io.close
+  end
+  def each
+    until @io.eof? do
+      leader = @io.read(24)
+      record = Marc4R::Record.new(leader)
+      record_body = @io.read(record.leader.record_length - 24)
+      load_record(record, record_body)
+      yield record
+    end
+  end
+  def load_record(record, content)
+    leader = record.leader
+    dir_length = leader.data_base_address - 25
+    raise Marc4R::MarcException.new("invalid directory length #{dir_length}\n#{leader.to_s}\n#{content}") unless (dir_length % 12) == 0
+    size = dir_length / 12
+
+    # if MARC 21 then check encoding
+    unless @encoding
+      if leader.char_coding_scheme.eql? ' '
+        @encoding = Encoding.find("ISO-8859-1")
+      elsif leader.char_coding_scheme.eql? 'a'
+        @encoding = Encoding.find("UTF-8")
+      end
+    end
+    
+    tags = []
+    lengths = []
+    starts = []
+    offset = 0
+    size.times do
+      tags << content.slice(offset,3)
+      lengths << content.slice(offset+3,4).to_i
+      offset += 12
+    end
+
+    unless content[offset] == Marc4R::Terminators::FIELD_TERMINATOR
+      raise Marc4R::MarcException.new("expected field terminator at end of directory")
+    else
+      offset += 1
+    end
+
+    (0...size).each do |i|
+      field_content = content.slice(offset,lengths[i])
+      offset += lengths[i]
+      unless field_content[-1] == Marc4R::Terminators::FIELD_TERMINATOR
+        raise Marc4R::MarcException.new("expected field terminator at end of field #{field_content}")
+      end
+      if Marc4R::Fields::ControlField.accepts_tag(tags[i])
+        data = field_content[0...-1].force_encoding(@encoding)
+        record.control_fields << Marc4R::Fields::ControlField.new(tags[i], data)
+      else # read a data field
+        record.data_fields << data_field(tags[i],field_content[0...-1])
+      end
+    end
+    unless content[offset] == Marc4R::Terminators::RECORD_TERMINATOR
+      raise Marc4R::MarcException.new("expected record terminator")
+    end
+  end
+  def data_field(tag, field_content)
+    raise Marc4R::MarcException.new("malformed data data field content") if field_content.size < 2
+    ind1 = field_content[0]
+    ind2 = field_content[1]
+
+    data_field = Marc4R::Fields::DataField.new(tag, ind1, ind2)
+    if field_content[2] == Marc4R::Terminators::SUBFIELD_TERMINATOR
+      subfields = field_content[3..-1].split(Marc4R::Terminators::SUBFIELD_TERMINATOR)
+      subfields = subfields.collect {|x| [x[0], x.slice(1..-1).force_encoding(@encoding)]}
+      subfields.flatten!
+    else
+      subfields = []
+    end
+    return Marc4R::Fields::DataField.new(tag, ind1, ind2, *subfields)
+  end
+end

data/lib/marc4r/record.rb

@@ -0,0 +1,92 @@
+class Marc4R::Record
+
+  attr_accessor :leader, :control_fields ,:data_fields
+
+  def initialize(leader = nil)
+    self.leader = Leader.new(leader)
+    self.control_fields = []
+    self.data_fields = []
+  end
+
+  def control_number
+    control_fields('001').first.data
+  end
+
+  def control_number_field
+    control_fields('001').first
+  end
+
+  def control_fields(*tags)
+    tags.empty? ? @control_fields : @control_fields.select {|field| tags.include? field.tag}
+  end
+
+  def data_fields(*tags)
+    tags.empty? ? @data_fields : @data_fields.select {|field| tags.include? field.tag}
+  end
+
+  def find(*args)
+
+    pattern = (args.last.is_a? Regexp) ? args.pop : nil
+    fields = (control_fields(*args) + data_fields(*args))
+    if pattern
+      fields = fields.select {|x| x =~ pattern}
+    end
+    fields
+  end
+
+  class Leader
+    attr_accessor :id, :record_length, :record_status, :record_type,
+                  :impl_defined1, :char_coding_scheme, :indicator_count,
+                  :subfield_code_count, :data_base_address, :impl_defined2, :entry_map
+
+    def initialize(value = nil)
+      parse(value) if value
+    end
+    def eql? x
+      (x.is_a? Marc4R::Record::Leader) && self.to_s.eql?(x.to_s)
+    end
+    private
+    def parse(value)
+      raise Marc4R::MarcException.new("leader should be 24 characters, was #{value.size}") unless value.size == 24
+    # The logical record length (Position 0-4).
+      @record_length = value[0..4].to_i
+    # The record status (Position 5).
+      @record_status = value[5]
+    # Type of record (Position 6).
+      @record_type = value[6]
+    # Implementation defined (Position 7-8).
+      @impl_defined1 = [value[7],value[8]]
+    # Character coding scheme (Position 9).
+      @char_coding_scheme = value[9]
+    # The indicator count (Position 10).
+      @indicator_count = value[10].to_i
+    # The subfield code length (Position 11).
+      @subfield_code_count = value[11].to_i
+    # The base address of data (Position 12-16).
+      @data_base_address = value[12..16].to_i
+    # Implementation defined (Position 17-18).
+      @impl_defined2 = [value[17],value[18]]
+    # Entry map (Position 19-23).
+      @entry_map = (19..23).collect {|x| value[x]}
+    end
+    def zero_pad(int, min_length)
+      log = (int > 0) ? Math.log10(int).ceil : 1
+      (log < min_length) ? (0...min_length-log).inject(int.to_s) {|m,v| m.prepend '0'}
+                         : int.to_s
+    end
+    public
+    def to_s
+      strings = @record_length ? [zero_pad(@record_length,5)] : []
+      strings << @record_status if @record_status
+      strings << @record_type if @record_type
+      strings += @impl_defined1 if @impl_defined1
+      strings << @char_coding_scheme if @char_coding_scheme
+      strings << @indicator_count.to_s if @indicator_count
+      strings << @subfield_code_count.to_s if @subfield_code_count
+      strings << zero_pad(@data_base_address,5) if @data_base_address
+      strings += @impl_defined2 if @impl_defined2
+      strings += @entry_map if @entry_map
+      strings.join('')
+    end
+  end
+end

data/marc4r.gemspec

@@ -0,0 +1,21 @@
+# -*- encoding: utf-8 -*-
+version = File.read(File.expand_path("../VERSION",__FILE__)).strip
+
+Gem::Specification.new do |gem|
+  gem.authors       = ["barmintor"]
+  gem.email         = ["armintor@gmail.com"]
+  gem.description   = %q{MARC4R is a Ruby MARC parser based on MARC4J}
+  gem.summary       = %q{MARC4R is a Ruby MARC parser based on MARC4J}
+  gem.homepage      = "http://github.com/barmintor/marc4r"
+
+  gem.files         = `git ls-files`.split($\)
+  gem.executables   = gem.files.grep(%r{^bin/}).map{ |f| File.basename(f) }
+  gem.test_files    = gem.files.grep(%r{^(test|spec|features)/})
+  gem.name          = "marc4r"
+  gem.require_paths = ["lib"]
+  gem.version       = version
+  gem.license       = 'MIT'
+
+  gem.add_development_dependency 'rake'
+  gem.add_development_dependency 'rspec-its'
+end

data/spec/fixtures/OCLC_814388508.xml

@@ -0,0 +1,123 @@
+<?xml version="1.0" encoding="UTF-8"?><collection xmlns="http://www.loc.gov/MARC21/slim">
+  <record>
+    <leader>02219cam a2200385Ia 4500</leader>
+    <controlfield tag="001">ocn814388508</controlfield>
+    <controlfield tag="003">OCoLC</controlfield>
+    <controlfield tag="005">20140115103422.0</controlfield>
+    <controlfield tag="008">121025s2011    gr ab    b    100 0 gre d</controlfield>
+    <datafield tag="040" ind1=" " ind2=" ">
+      <subfield code="a">JPG</subfield>
+      <subfield code="c">JPG</subfield>
+      <subfield code="d">JPG</subfield>
+    </datafield>
+    <datafield tag="066" ind1=" " ind2=" ">
+      <subfield code="c">(S</subfield>
+    </datafield>
+    <datafield tag="020" ind1=" " ind2=" ">
+      <subfield code="a">9789608838765</subfield>
+    </datafield>
+    <datafield tag="020" ind1=" " ind2=" ">
+      <subfield code="a">9608838762</subfield>
+    </datafield>
+    <datafield tag="035" ind1=" " ind2=" ">
+      <subfield code="a">(OCoLC)814388508</subfield>
+    </datafield>
+    <datafield tag="041" ind1="0" ind2=" ">
+      <subfield code="a">gre</subfield>
+      <subfield code="a">eng</subfield>
+      <subfield code="a">ger</subfield>
+      <subfield code="b">eng</subfield>
+      <subfield code="b">gre</subfield>
+    </datafield>
+    <datafield tag="043" ind1=" " ind2=" ">
+      <subfield code="a">mm-----</subfield>
+    </datafield>
+    <datafield tag="050" ind1=" " ind2="4">
+      <subfield code="a">NA9070</subfield>
+      <subfield code="b">.A36 2011</subfield>
+    </datafield>
+    <datafield tag="049" ind1=" " ind2=" ">
+      <subfield code="a">VA@@</subfield>
+    </datafield>
+    <datafield tag="245" ind1="0" ind2="3">
+      <subfield code="6">880-01</subfield>
+      <subfield code="a">Hē agora stē Mesogeio :</subfield>
+      <subfield code="b">apo tous Homērikous heōs tous Rōmaikous chronous : Diethnes epistēmoniko synedrio, Kōs, 14-17 Apriliou 2011 = The agora in the Mediterranean : from Homeric to Roman times : International conference, Kos, 14-17 April 2011 /</subfield>
+      <subfield code="c">epistēmonikē epimeleia, Angelikē Giannikourē.</subfield>
+    </datafield>
+    <datafield tag="246" ind1="3" ind2="1">
+      <subfield code="a">Agora in the Mediterranean</subfield>
+    </datafield>
+    <datafield tag="260" ind1=" " ind2=" ">
+      <subfield code="6">880-02</subfield>
+      <subfield code="a">Athēna :</subfield>
+      <subfield code="b">Hypourgeio Politismou kai Tourismou, Archaiologiko Instituto Aigaiakōn Spoudōn,</subfield>
+      <subfield code="c">2011.</subfield>
+    </datafield>
+    <datafield tag="300" ind1=" " ind2=" ">
+      <subfield code="a">438 p. :</subfield>
+      <subfield code="b">ill., maps ;</subfield>
+      <subfield code="c">28 cm.</subfield>
+    </datafield>
+    <datafield tag="504" ind1=" " ind2=" ">
+      <subfield code="a">Includes bibliographical references.</subfield>
+    </datafield>
+    <datafield tag="546" ind1=" " ind2=" ">
+      <subfield code="a">Contributions in Greek or English, with abstracts in the other language; 1 contribution in German, with abstract in Greek.</subfield>
+    </datafield>
+    <datafield tag="650" ind1=" " ind2="0">
+      <subfield code="a">Plazas</subfield>
+      <subfield code="z">Mediterranean Region</subfield>
+      <subfield code="v">Congresses.</subfield>
+    </datafield>
+    <datafield tag="650" ind1=" " ind2="0">
+      <subfield code="a">Cities and towns, Ancient</subfield>
+      <subfield code="z">Mediterranean Region</subfield>
+      <subfield code="v">Congresses.</subfield>
+    </datafield>
+    <datafield tag="650" ind1=" " ind2="0">
+      <subfield code="a">Excavations (Archaeology)</subfield>
+      <subfield code="z">Mediterranean Region</subfield>
+      <subfield code="v">Congresses.</subfield>
+    </datafield>
+    <datafield tag="651" ind1=" " ind2="0">
+      <subfield code="a">Mediterranean Region</subfield>
+      <subfield code="x">Antiquities</subfield>
+      <subfield code="v">Congresses.</subfield>
+    </datafield>
+    <datafield tag="700" ind1="1" ind2=" ">
+      <subfield code="6">880-03</subfield>
+      <subfield code="a">Giannikourē, A.</subfield>
+    </datafield>
+    <datafield tag="710" ind1="1" ind2=" ">
+      <subfield code="6">880-04</subfield>
+      <subfield code="a">Greece.</subfield>
+      <subfield code="b">Archaiologiko Institouto Aigaiakōn Spoudōn.</subfield>
+    </datafield>
+    <datafield tag="880" ind1="0" ind2="2">
+      <subfield code="6">245-01/(S</subfield>
+      <subfield code="a">Η αγορά στη Μεσόγειο :</subfield>
+      <subfield code="b">από τους Ομηρικούς έως τους Ρωμαϊκούς χρόνους : Διεθνές επιστημονικό συνέδριο, Κως, 14-17 Απριλίου 2011 = The agora in the Mediterranean : from Homeric to Roman times : International conference, Kos, 14-17 April 2011 /</subfield>
+      <subfield code="c">επιστημονικη επιμέλεια Αγγελική Γιαννικουρη.</subfield>
+    </datafield>
+    <datafield tag="880" ind1=" " ind2=" ">
+      <subfield code="6">260-02/(S</subfield>
+      <subfield code="a">Αθηνα :</subfield>
+      <subfield code="b">Υπουργείο Πολιτισμου και Τουρισμου, Αρχαιολογικου Ινστιτουτου Αιγαιακών Σπουδών,</subfield>
+      <subfield code="c">2011.</subfield>
+    </datafield>
+    <datafield tag="880" ind1="1" ind2=" ">
+      <subfield code="6">700-03/(S</subfield>
+      <subfield code="a">Γιαννικουρη, Α.</subfield>
+    </datafield>
+    <datafield tag="880" ind1="1" ind2=" ">
+      <subfield code="6">710-04/(S</subfield>
+      <subfield code="a">Greece.</subfield>
+      <subfield code="b">Αρχαιολογικου Ινστιτουτου Αιγαιακών Σπουδών.</subfield>
+    </datafield>
+    <datafield tag="994" ind1=" " ind2=" ">
+      <subfield code="a">C0</subfield>
+      <subfield code="b">VA@</subfield>
+    </datafield>
+  </record>
+</collection>