mab2 0.0.2 → 0.0.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: cbecb44d9edd1e649510ee35cc401277ba3cff9d
-  data.tar.gz: 9aabd361f1ba9b69727b08435176d8a2e9c37d37
+  metadata.gz: d1659413e48d57b2b853237962d2763d20bf6288
+  data.tar.gz: cd5bbecc8442e0769c27539e40283b41e361dc96
 SHA512:
-  metadata.gz: a3f6292cd094c7a5834cd34e63e5983b9f47de5eef2520f1d47dc8dc4f6da073545052b2b570e7e36892c467dba377694cac7756eb8edbe00437d9e26907b680
-  data.tar.gz: e696b3d6612c4f1840db905eed35a0873de66c2c684806e1f5e245b2eac55b436f5ba0c80b86e8d59c0cc89887aea1c64d8bc5682c7a7641e6019283ca5cbd44
+  metadata.gz: bfde3b1e18c169f751db9faf12fa988c614ef63dcd76600ea9b143a1e9ecc95deebbce608579843b80dbdb3c517d1abfd7f242d7468294a2ea24108bcd684d37
+  data.tar.gz: 54e634c0052d6594bb3cc6e8920d2452baf901f5250bc8de34d0611088a222fb36ca83c63fbe060946d828c640a60b79ac42ba5d16111caf1413c304bdf1671d

data/lib/mab2/aleph_sequential_adapter.rb

@@ -0,0 +1,10 @@
+class Mab2::AlephSequentialAdapter
+
+  attr_accessor :controlfields, :datafields
+
+  def initialize(aleph_sequential)
+    @controlfields = {}
+    @datafields = {}
+  end
+
+end

data/lib/mab2/document/datafield.rb

@@ -0,0 +1,15 @@
+class Mab2::Document::Datafield
+
+  attr_accessor :indicators, :subfields, :tag
+
+  def initialize(tag, indicators, subfields)
+    @tag = tag
+    @indicators = indicators
+    @subfields = subfields.map { |subfield| Mab2::Document::Subfield.new(subfield) }
+  end
+
+  def get_subfield(code)
+    @subfields.find { |subfield| subfield.name == code.to_sym }
+  end
+
+end

data/lib/mab2/document/scope.rb

@@ -0,0 +1,89 @@
+class Mab2::Document::Scope
+
+  attr_accessor :scope_subfields
+
+  def initialize(document)
+    @document = document
+  end
+
+  def controlfield(tag)
+    @document.controlfields[tag.to_sym].try(:chars) || []
+  end
+
+  def datafield(tag, indicators)
+    @scope_datafields = []
+    indicators = sanitize_indicators(indicators)
+
+    selected_datafields = if !(datafields_by_tag = @document.datafields[tag.to_sym]).nil?
+      datafields_by_tag.select do |indicator_one_key, indicator_one_value|
+        indicators[:ind1].nil? || (indicators[:ind1].any? { |ind1| ind1.length > 1 && ind1.to_s.starts_with?('-') } ^ indicators[:ind1].include?(indicator_one_key)) && indicator_one_value.any? do |indicator_two_key, indicator_two_value|
+          indicators[:ind2].nil? || indicators[:ind2].include?(indicator_two_key)
+        end
+      end
+    end
+
+    if !selected_datafields.nil?
+      selected_datafields.each_pair do |indicator_one_key, indicator_one_value|
+        indicator_one_value.each_pair do |indicator_two_key, indicator_two_value|
+          indicator_two_value.each do |subfields|
+            @scope_datafields.push Mab2::Document::Datafield.new(tag, indicators, subfields)
+          end
+        end
+      end
+    end
+
+    self
+  end
+
+  def get
+    self
+  end
+
+  def fields
+    @scope_datafields
+  end
+
+  def scope_subfields
+    @scope_subfields || @scope_datafields.map { |datafield| datafield.subfields }.flatten.presence
+  end
+
+  def subfield(subfield_codes)
+    # make subfield_codes an array (always)
+    subfield_codes = subfield_codes.is_a?(Array) ? subfield_codes : [subfield_codes]
+
+    @scope_subfields = []
+
+    @scope_datafields.each do |datafield|
+      @scope_subfields.push datafield.subfields.select { |subfield| subfield_codes.include? subfield.name.to_s }
+    end
+
+    @scope_subfields.flatten!
+
+    self
+  end
+
+  def subfields
+    scope_subfields
+  end
+
+  def value(options = {})
+    if !(_values = values(options)).nil?
+      _values.first
+    end
+  end
+
+  def values(options = {})
+    if !scope_subfields.nil?
+      subfield_values = scope_subfields.map { |subfield| subfield.value }
+      @scope_values = (join_string = options[:join_subfields]) ? subfield_values.join(join_string == true ? '' : join_string) : subfield_values
+    end
+  end
+
+ private
+
+  # ensure a hash with array values
+  def sanitize_indicators(indicators)
+    indicators.inject({}) { |hash, (k, v)| hash[k] = (v.is_a?(Array) ? v.map(&:to_sym) : [v.to_sym]); hash }
+  end
+
+end

data/lib/mab2/document/subfield.rb

@@ -0,0 +1,10 @@
+class Mab2::Document::Subfield
+
+  attr_accessor :name, :value
+
+  def initialize(hash)
+    @name = hash.keys.first
+    @value = hash.values.first
+  end
+
+end

data/lib/mab2/document.rb

@@ -1,28 +1,36 @@
-require 'mab2/document_scope'
-require 'nokogiri'
+require 'mab2/aleph_sequential_adapter'
+require 'mab2/mab_xml_adapter'
 
-module Mab2
-  class Document
+class Mab2::Document
+  require_relative 'document/datafield'
+  require_relative 'document/scope'
+  require_relative 'document/subfield'
 
-    attr_accessor :nokogiri_document
+  attr_accessor :controlfields
+  attr_accessor :datafields
 
-    # we assume someone calls us like this Mab2::Document.new('<xml>...'</xml>')
-    def initialize(xml)
-      @nokogiri_document = Nokogiri::XML(xml).remove_namespaces!.css('OAI-PMH ListRecords record metadata record')
+  def initialize(input)
+    mab = if input.is_a?(Mab2::MabXmlAdapter) || input.is_a?(Mab2::AlephSequentialAdapter)
+      input
+    elsif input.is_a?(String) && input.starts_with?('<?xml')
+      Mab2::MabXmlAdapter.new(input)
+    elsif input.is_a?(String) && input.starts_with?('LEADER')
+      Mab2::AlephSequentialAdapter.new(input)
+    else
+      raise 'Unknown input format or adapter!'
     end
 
-    def controlfield(tag)
-      Mab2::DocumentScope.new(self).controlfield(tag)
-    end
-
-    def datafield(tag, options = {})
-      Mab2::DocumentScope.new(self).datafield(tag, options)
-    end
+    @controlfields = mab.controlfields
+    @datafields = mab.datafields
+  end
 
-    # mabmapper uses field instead of datafield
-    def field(tag, options = {})
-      datafield(tag, options)
-    end
+  def controlfield(tag)
+    Mab2::Document::Scope.new(self).controlfield(tag)
+  end
 
+  # mabmapper uses field instead of datafield
+  def field(tag, options = {})
+    Mab2::Document::Scope.new(self).datafield(tag, options)
   end
+
 end

data/lib/mab2/mab_xml_adapter.rb

@@ -0,0 +1,50 @@
+require 'nokogiri'
+
+class Mab2::MabXmlAdapter
+
+  attr_accessor :controlfields, :datafields
+
+  def initialize(xml)
+    record = Nokogiri::XML(xml).remove_namespaces!.xpath('//OAI-PMH/ListRecords/record/metadata/record')
+
+    @controlfields = controlfields_from_mabxml(record)
+    @datafields = datafields_from_mabxml(record)
+  end
+
+ private
+
+  def controlfields_from_mabxml(mabxml)
+    if mabxml.is_a? Nokogiri::XML::NodeSet
+      mabxml.xpath('//controlfield').inject({}) do |controlfields, xml_node|
+        controlfields[xml_node.attr('tag').to_sym] = xml_node.content
+
+        # return resulting hash for inject's sake
+        controlfields
+      end
+    end
+  end
+
+  def datafields_from_mabxml(mabxml)
+    if mabxml.is_a? Nokogiri::XML::NodeSet
+      mabxml.xpath('//datafield').inject({}) do |datafields, xml_node|
+        tag = xml_node.attr('tag').to_sym
+        ind1 = xml_node.attr('ind1').to_sym
+        ind2 =  xml_node.attr('ind2').to_sym
+
+        existing_subfields = datafields[tag].try(:[], ind1).try(:[], ind2)
+
+        datafields[tag] = {
+          ind1 => {
+            ind2 => (existing_subfields.nil? ? [] : existing_subfields).push(xml_node.xpath('subfield').map do |subfield|
+              { subfield.attr('code').to_sym => subfield.content }
+            end)
+          }
+        }
+
+        # return resulting hash for inject's sake
+        datafields
+      end
+    end
+  end
+
+end

data/lib/mab2/version.rb

@@ -1,3 +1,3 @@
 module Mab2
-  VERSION = "0.0.2"
+  VERSION = "0.0.3"
 end

data/lib/mab2.rb

@@ -1,2 +1,4 @@
-require 'mab2/document'
-require 'mab2/version'
+module Mab2
+  require 'mab2/document'
+  require 'mab2/version'
+end

data/mab2.gemspec

@@ -20,7 +20,7 @@ Gem::Specification.new do |spec|
 
   spec.add_dependency 'nokogiri', '~> 1.6.0'
 
-  spec.add_development_dependency 'activesupport', '>= 3.2.0', '< 5.0.0'
+  spec.add_development_dependency 'activesupport', '>= 3.2.0', '< 4.0.0'
   spec.add_development_dependency 'bundler',       '~> 1.3'
   spec.add_development_dependency 'minitest',      '~> 5.0.6'
   spec.add_development_dependency 'pry',           '~> 0.9.12.2'

data/test/mab2/test_controlfield.rb

@@ -15,4 +15,10 @@ module TestControlfield
     assert_equal doc.controlfield('050').at(3), 'c'
   end
 
+  # accessing non-existing controlfields
+  def test_controlfield_4
+    doc = Mab2::Document.new(load_mab('PAD01.000806191.PRIMO.xml'))
+    assert_equal doc.controlfield('052').at(0), nil
+  end
+
 end

data/test/mab2/test_datafield.rb

@@ -1,21 +1,131 @@
 module TestDatafield
 
+  #
+  # Synthetic dsl tests
+  #
+  # full blown path, single value
+  def test_dsl_1
+    doc = Mab2::Document.new(load_mab('dsl_1.xml'))
+    assert_equal 'Some string', doc.field('100', ind1: '1', ind2: '2').subfield('a').get.value
+  end
+
+  # missing ind2, single value
+  def test_dsl_1a
+    doc = Mab2::Document.new(load_mab('dsl_1.xml'))
+    assert_equal 'Some string', doc.field('100', ind1: '1').subfield('a').get.value
+  end
+
+  # missing ind1 + ind2, single value
+  def test_dsl_1b
+    doc = Mab2::Document.new(load_mab('dsl_1.xml'))
+    assert_equal 'Some string', doc.field('100').subfield('a').get.value
+  end
+
+  # missing ind1 + ind2, missing subfield, single value
+  def test_dsl_1c
+    doc = Mab2::Document.new(load_mab('dsl_1.xml'))
+    assert_equal 'Some string', doc.field('100').get.value
+  end
+
+  # missing ind1 + ind2, missing subfield, no get, single value
+  def test_dsl_1d
+    doc = Mab2::Document.new(load_mab('dsl_1.xml'))
+    assert_equal 'Some string', doc.field('100').value
+  end
+
+  # missing ind1 + ind2, missing subfield, no get, explicite values
+  def test_dsl_1e
+    doc = Mab2::Document.new(load_mab('dsl_1.xml'))
+    assert_equal ['Some string'], doc.field('100').values
+  end
+
+  # missing ind1 + ind2, missing subfield, get fields
+  def test_dsl_1f
+    doc = Mab2::Document.new(load_mab('dsl_1.xml'))
+    fields = doc.field('100').get.fields
+    assert_equal 'Some string', fields.first.get_subfield('a').value
+  end
+
+  # multiple datafields with same tag
+  def test_dsl_2
+    doc = Mab2::Document.new(load_mab('dsl_2.xml'))
+    assert_equal ["Politik", "Theorie", "(DE-588)4046563-9", "Politiktheorie", "(DE-588)4046563-9", "Politische Theorien", "(DE-588)4046563-9", "Politik", "Theorie", "(DE-588)4046563-9", "Politiktheorie", "(DE-588)4046563-9", "Politische Theorien", "(DE-588)4046563-9"], doc.field('PSW', ind1: '-', ind2: '1').get.values
+  end
+
+  # multiple datafields with same tag, filtered by subfield
+  def test_dsl_2a
+    doc = Mab2::Document.new(load_mab('dsl_2.xml'))
+    assert_equal ["Politik", "Politiktheorie", "Politische Theorien", "Politik", "Politiktheorie", "Politische Theorien"], doc.field('PSW', ind1: '-', ind2: '1').subfield('s').get.values
+  end
+
+  # multiple datafields with same tag, filtered by subfield only existent in two of n datafields
+  def test_dsl_2b
+    doc = Mab2::Document.new(load_mab('dsl_2.xml'))
+    assert_equal ["Theorie", "Theorie"], doc.field('PSW', ind1: '-', ind2: '1').subfield('x').get.values
+  end
+
+  # multiple datafields with same tag, filtered by subfield only existent in two of n datafields, only first value
+  def test_dsl_2c
+    doc = Mab2::Document.new(load_mab('dsl_2.xml'))
+    assert_equal "Theorie", doc.field('PSW', ind1: '-', ind2: '1').subfield('x').get.value
+  end
+
+  #
+  # Real world mab files
+  #
+
   # ind1 missing
   def test_datafield_1
     doc = Mab2::Document.new(load_mab('PAD01.000806191.PRIMO.xml'))
     assert_equal '20000217', doc.field('LOC', ind2: ' ').subfield('k').get.value
   end
 
-  # ind1 missing, first value of multi ind2/subfield query
+  # ind1 missing, multi ind2/subfield, first value
   def test_datafield_2
     doc = Mab2::Document.new(load_mab('PAD01.000806191.PRIMO.xml'))
     assert_equal 'Gockel, Simone', doc.field("100", ind2: ['1', '2']).subfield(['a','p','c','n','b']).get.value
   end
 
+  # ind1/ind2 missing, all values
   def test_datafield_3
     doc = Mab2::Document.new(load_mab('PAD01.000978177.PRIMO.xml'))
     assert_equal ["Bajme, Klaus", "Beyme, Klaus", "Beyme, K. <<von>>", "Bajme, Klaus", "Beyme, Klaus", "Beyme, K. <<von>>"], doc.field('PPE').subfield(['a', 'p']).get.values
   end
 
+  # ind1 negated '-z' => all but 'z', all values
+  def test_datafield_4
+    doc = Mab2::Document.new(load_mab('PAD01.000978177.PRIMO.xml'))
+    assert_equal ["978-3-518-28569-5"], doc.field('540', ind1: '-z', ind2: '1').subfield('a').get.values
+  end
+
+  # simple fields
+  def test_datafield_5
+    doc = Mab2::Document.new(load_mab('PAD01.000978177.PRIMO.xml'))
+    fields = doc.field('655').subfield(['u', '3', 'z', 't']).get.fields
+    assert_equal 'http://digitool.hbz-nrw.de:1801/webclient/DeliveryManager?pid=2280738&custom_att_2=simple_viewer', fields.first.get_subfield('u').value
+  end
+
+  # iterating over fields
+  def test_datafield_5
+    doc = Mab2::Document.new(load_mab('PAD01.000978177.PRIMO.xml'))
+    fields = doc.field('655').subfield(['u', '3', 'z', 't']).get.fields
+    fields.each do |field|
+      assert_equal 'VIEW', field.get_subfield('t').value
+      assert_equal 'http://digitool.hbz-nrw.de:1801/webclient/DeliveryManager?pid=2280738&custom_att_2=simple_viewer', field.get_subfield('u').value
+    end
+  end
+
+  #def test_datafield_6
+  #  doc = Mab2::Document.new(load_mab('PAD01.000978177.PRIMO.xml'))
+  #  fields = doc.field('PSW').subfield(['s', '9']).get.fields
+  #  binding.pry
+  #end
+
+  # join_subfields
+  #def test_datafield_4
+  #  doc = Mab2::Document.new(load_mab('PAD01.000978177.PRIMO.xml'))
+  #  w  = doc.field('525', ind2: '1').subfield(['p','a']).get.value(join_subfields: ': ')
+  #  binding.pry
+  #end
 
 end

data/test/mab_files/dsl_1.xml

@@ -0,0 +1,14 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<OAI-PMH xmlns="http://www.openarchives.org/OAI/2.0/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.openarchives.org/OAI/2.0/ http://www.openarchives.org/OAI/2.0/OAI-PMH.xsd">
+  <ListRecords>
+    <record>
+      <metadata>
+        <record xmlns="http://www.ddb.de/professionell/mabxml/mabxml-1.xsd">
+          <datafield tag="100" ind1="1" ind2="2">
+            <subfield code="a">Some string</subfield>
+          </datafield>
+        </record>
+      </metadata>
+    </record>
+  </ListRecords>
+</OAI-PMH>

data/test/mab_files/dsl_2.xml

@@ -0,0 +1,37 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<OAI-PMH xmlns="http://www.openarchives.org/OAI/2.0/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.openarchives.org/OAI/2.0/ http://www.openarchives.org/OAI/2.0/OAI-PMH.xsd">
+  <ListRecords>
+    <record>
+      <metadata>
+        <record xmlns="http://www.ddb.de/professionell/mabxml/mabxml-1.xsd">
+          <datafield tag="PSW" ind1="-" ind2="1">
+            <subfield code="s">Politik</subfield>
+            <subfield code="x">Theorie</subfield>
+            <subfield code="9">(DE-588)4046563-9</subfield>
+          </datafield>
+          <datafield tag="PSW" ind1="-" ind2="1">
+            <subfield code="s">Politiktheorie</subfield>
+            <subfield code="9">(DE-588)4046563-9</subfield>
+          </datafield>
+          <datafield tag="PSW" ind1="-" ind2="1">
+            <subfield code="s">Politische Theorien</subfield>
+            <subfield code="9">(DE-588)4046563-9</subfield>
+          </datafield>
+          <datafield tag="PSW" ind1="-" ind2="1">
+            <subfield code="s">Politik</subfield>
+            <subfield code="x">Theorie</subfield>
+            <subfield code="9">(DE-588)4046563-9</subfield>
+          </datafield>
+          <datafield tag="PSW" ind1="-" ind2="1">
+            <subfield code="s">Politiktheorie</subfield>
+            <subfield code="9">(DE-588)4046563-9</subfield>
+          </datafield>
+          <datafield tag="PSW" ind1="-" ind2="1">
+            <subfield code="s">Politische Theorien</subfield>
+            <subfield code="9">(DE-588)4046563-9</subfield>
+          </datafield>
+        </record>
+      </metadata>
+    </record>
+  </ListRecords>
+</OAI-PMH>

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: mab2
 version: !ruby/object:Gem::Version
-  version: 0.0.2
+  version: 0.0.3
 platform: ruby
 authors:
 - Michael Sievers
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2013-09-25 00:00:00.000000000 Z
+date: 2013-11-19 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: nokogiri
@@ -33,7 +33,7 @@ dependencies:
         version: 3.2.0
     - - <
       - !ruby/object:Gem::Version
-        version: 5.0.0
+        version: 4.0.0
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
@@ -43,7 +43,7 @@ dependencies:
         version: 3.2.0
     - - <
       - !ruby/object:Gem::Version
-        version: 5.0.0
+        version: 4.0.0
 - !ruby/object:Gem::Dependency
   name: bundler
   requirement: !ruby/object:Gem::Requirement
@@ -127,14 +127,20 @@ files:
 - README.md
 - Rakefile
 - lib/mab2.rb
+- lib/mab2/aleph_sequential_adapter.rb
 - lib/mab2/document.rb
-- lib/mab2/document_scope.rb
+- lib/mab2/document/datafield.rb
+- lib/mab2/document/scope.rb
+- lib/mab2/document/subfield.rb
+- lib/mab2/mab_xml_adapter.rb
 - lib/mab2/version.rb
 - mab2.gemspec
 - test/mab2/test_controlfield.rb
 - test/mab2/test_datafield.rb
 - test/mab_files/PAD01.000806191.PRIMO.xml
 - test/mab_files/PAD01.000978177.PRIMO.xml
+- test/mab_files/dsl_1.xml
+- test/mab_files/dsl_2.xml
 - test/test_helper.rb
 - test/test_mab2.rb
 homepage: https://github.com/msievers/mab2
@@ -166,5 +172,7 @@ test_files:
 - test/mab2/test_datafield.rb
 - test/mab_files/PAD01.000806191.PRIMO.xml
 - test/mab_files/PAD01.000978177.PRIMO.xml
+- test/mab_files/dsl_1.xml
+- test/mab_files/dsl_2.xml
 - test/test_helper.rb
 - test/test_mab2.rb

data/lib/mab2/document_scope.rb

@@ -1,64 +0,0 @@
-class Mab2::DocumentScope
-
-  def initialize(scoped_document)
-    @doc = scoped_document.nokogiri_document
-  end
-
-  def controlfield(tag)
-    @doc.css("controlfield[tag='#{tag}']").text.chars
-  end
-
-  def datafield(tag, indicators)
-    @tag = tag
-    @indicators = indicators.inject({}) { |hash, (k, v)| hash[k] = (v.is_a?(Array) ? v : [v]); hash }
-    self
-  end
-
-  def get
-    self
-  end
-
-  def subfield(subfield_codes)
-    @subfield_codes = subfield_codes.is_a?(Array) ? subfield_codes : [subfield_codes]
-    self
-  end
-
-  def value
-    values.first
-  end
-
-  def values
-    datafield_selectors = []
-    
-    if @indicators[:ind1].present? && @indicators[:ind2].blank?
-      @indicators[:ind1].each do |ind1_value|
-        datafield_selectors.push("datafield[tag='#{@tag}'][ind1='#{ind1_value}']")
-      end
-    elsif @indicators[:ind1].blank? && @indicators[:ind2].present?
-      @indicators[:ind2].each do |ind2_value|
-        datafield_selectors.push("datafield[tag='#{@tag}'][ind2='#{ind2_value}']")
-      end
-    elsif @indicators[:ind1].present? && @indicators[:ind2].present?
-      @indicators[:ind1].each do |ind1_value|
-        @indicators[:ind2].each do |ind2_value|
-          datafield_selectors.push("datafield[tag='#{@tag}'][ind1='#{ind1_value}'][ind2='#{ind2_value}']")
-        end
-      end
-    else
-      datafield_selectors.push("datafield[tag='#{@tag}']")
-    end
-
-    css_selector = if @subfield_codes.present?
-      datafield_selectors.map do |selector|
-        @subfield_codes.map do |subfield_code|
-          "#{selector} subfield[code='#{subfield_code}']"
-        end
-      end
-    else
-      indicator_selectors
-    end.flatten.join(', ')
-
-    @doc.css(css_selector).map(&:text)
-  end
-
-end