marc 1.0.4 → 1.2.0

data/test/marc8/tc_to_unicode.rb

@@ -1,15 +1,12 @@
-# encoding: UTF-8
+require "test/unit"
+require "marc"
 
-require 'test/unit'
-require 'marc'
+require "marc/marc8/to_unicode"
 
-require 'marc/marc8/to_unicode'
-
-require 'unf'
+require "unf"
 
 if "".respond_to?(:encoding)
 
-
   class TestMarc8ToUnicode < Test::Unit::TestCase
     def test_empty_string
       value = MARC::Marc8::ToUnicode.new.transcode("")
@@ -35,18 +32,18 @@ if "".respond_to?(:encoding)
       # two data files, marc8 and utf8, with line-by-line correspondences.
       #
       # For now, we have NOT included proprietary III encodings in our test data!
-      utf8_file   = File.open( File.expand_path("../data/test_utf8.txt", __FILE__), "r:UTF-8")
-      marc8_file  = File.open( File.expand_path("../data/test_marc8.txt", __FILE__), "r:binary")
+      utf8_file = File.open(File.expand_path("../data/test_utf8.txt", __FILE__), "r:UTF-8")
+      marc8_file = File.open(File.expand_path("../data/test_marc8.txt", __FILE__), "r:binary")
 
       i = 0
       converter = MARC::Marc8::ToUnicode.new
 
       begin
-        while true do
+        loop do
           i += 1
 
-          utf8      = utf8_file.readline.chomp
-          marc8     = marc8_file.readline.chomp
+          utf8 = utf8_file.readline.chomp
+          marc8 = marc8_file.readline.chomp
 
           converted = converter.transcode(marc8)
 
@@ -55,11 +52,11 @@ if "".respond_to?(:encoding)
 
           assert_equal utf8, converted, "Test data line #{i}, expected converted to match provided utf8"
         end
-      rescue EOFError => each
+      rescue EOFError
         # just means the file was over, no biggie
         assert i > 1500, "Read as many lines as we expected to, at least 1500"
-      rescue Exception => e
-        $stderr.puts "Error at test data line #{i}"
+      rescue => e
+        warn "Error at test data line #{i}"
         raise e
       end
     end
@@ -67,23 +64,23 @@ if "".respond_to?(:encoding)
     def test_explicit_normalization
       # \xC1 is Marc8 "script small letter l", which under unicode
       # COMPAT normalization will turn into ordinary 'l'
-      marc8     = "Conversa\xF0c\xE4ao \xC1"
-      unicode   = "Conversação \u2113"
+      marc8 = "Conversa\xF0c\xE4ao \xC1"
+      unicode = "Conversação \u2113"
 
-      unicode_c   = UNF::Normalizer.normalize(unicode, :nfc)
-      unicode_kc  = UNF::Normalizer.normalize(unicode, :nfkc)
-      unicode_d  = UNF::Normalizer.normalize(unicode, :nfd)
-      unicode_kd  = UNF::Normalizer.normalize(unicode, :nfkd)
+      unicode_c = UNF::Normalizer.normalize(unicode, :nfc)
+      unicode_kc = UNF::Normalizer.normalize(unicode, :nfkc)
+      unicode_d = UNF::Normalizer.normalize(unicode, :nfd)
+      unicode_kd = UNF::Normalizer.normalize(unicode, :nfkd)
 
       converter = MARC::Marc8::ToUnicode.new
 
-      assert_equal unicode_c,  converter.transcode(marc8, :normalization => :nfc)
-      assert_equal unicode_kc, converter.transcode(marc8, :normalization => :nfkc)
-      assert_equal unicode_d,  converter.transcode(marc8, :normalization => :nfd)
-      assert_equal unicode_kd, converter.transcode(marc8, :normalization => :nfkd)
+      assert_equal unicode_c, converter.transcode(marc8, normalization: :nfc)
+      assert_equal unicode_kc, converter.transcode(marc8, normalization: :nfkc)
+      assert_equal unicode_d, converter.transcode(marc8, normalization: :nfd)
+      assert_equal unicode_kd, converter.transcode(marc8, normalization: :nfkd)
 
       # disable normalization for performance or something, we won't end up with NFC.
-      refute_equal unicode_c, converter.transcode(marc8, :normalization => nil)
+      refute_equal unicode_c, converter.transcode(marc8, normalization: nil)
     end
 
     def test_expand_ncr
@@ -91,7 +88,7 @@ if "".respond_to?(:encoding)
 
       marc8_ncr = "Weird &#x200F; &#xFFFD; but these aren't changed #x2000; &#200F etc."
       assert_equal "Weird \u200F \uFFFD but these aren't changed #x2000; &#200F etc.", converter.transcode(marc8_ncr)
-      assert_equal marc8_ncr, converter.transcode(marc8_ncr, :expand_ncr => false), "should not expand NCR if disabled"
+      assert_equal marc8_ncr, converter.transcode(marc8_ncr, expand_ncr: false), "should not expand NCR if disabled"
     end
 
     def test_bad_byte
@@ -130,7 +127,7 @@ if "".respond_to?(:encoding)
       converter = MARC::Marc8::ToUnicode.new
 
       bad_marc8 = "\e$1!PVK7oi$N!Q1!G4i$N!0p!Q+{6924f6}\e(B"
-      value = converter.transcode(bad_marc8, :invalid => :replace)
+      value = converter.transcode(bad_marc8, invalid: :replace)
 
       assert_equal "UTF-8", value.encoding.name
       assert value.valid_encoding?
@@ -145,7 +142,7 @@ if "".respond_to?(:encoding)
       converter = MARC::Marc8::ToUnicode.new
 
       bad_marc8 = "\e$1!PVK7oi$N!Q1!G4i$N!0p!Q+{6924f6}\e(B"
-      value = converter.transcode(bad_marc8, :invalid => :replace, :replace => "")
+      value = converter.transcode(bad_marc8, invalid: :replace, replace: "")
 
       assert_equal "UTF-8", value.encoding.name
       assert value.valid_encoding?
@@ -169,9 +166,8 @@ if "".respond_to?(:encoding)
 
       assert_equal "La Soci\u00E9t\x1B,", value
     end
-
   end
 else
-  require 'pathname'
-  $stderr.puts "\nTests not being run in ruby 1.9.x, skipping #{Pathname.new(__FILE__).basename}\n\n"
+  require "pathname"
+  warn "\nTests not being run in ruby 1.9.x, skipping #{Pathname.new(__FILE__).basename}\n\n"
 end

data/test/messed_up_leader.xml

@@ -0,0 +1,9 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<collection xmlns="http://www.loc.gov/MARC21/slim">
+  <record>
+    <leader>01301nam a22003618&lt; 4500</leader>
+    <controlfield tag="005">02-26-96</controlfield>
+    <controlfield tag="008">960213s1996    flu           000 0 eng  </controlfield>
+    <controlfield tag="001">991010422259705706</controlfield>
+    <datafield tag="010" ind1=" " ind2=" "><subfield code="a">95041713</subfield></datafield><datafield tag="020" ind1=" " ind2=" "><subfield code="a">1566700701</subfield><subfield code="q">(alk. paper)</subfield></datafield><datafield tag="020" ind1=" " ind2=" "><subfield code="a">9781566700702</subfield><subfield code="q">(alk. paper)</subfield></datafield><datafield tag="035" ind1=" " ind2=" "><subfield code="a">(NhD)b23774344-01dcl_inst</subfield></datafield><datafield tag="035" ind1=" " ind2=" "><subfield code="a">(CStRLIN)ocm33245123</subfield><subfield code="9">ExL</subfield></datafield><datafield tag="035" ind1=" " ind2=" "><subfield code="b">b23774344</subfield></datafield><datafield tag="040" ind1=" " ind2=" "><subfield code="a">DLC</subfield><subfield code="c">DLC</subfield><subfield code="d">NhD</subfield></datafield><datafield tag="050" ind1=" " ind2="4"><subfield code="a">QH545.A1</subfield><subfield code="b">R65 1996</subfield></datafield><datafield tag="100" ind1="1" ind2=" "><subfield code="a">Römbke, Jörg.</subfield><subfield code="0">http://id.loc.gov/authorities/names/n95093926.</subfield></datafield><datafield tag="245" ind1="1" ind2="0"><subfield code="a">Applied ecotoxicology /</subfield><subfield code="c">Jörg Römbke, Johann F. Moltmann.</subfield></datafield><datafield tag="260" ind1=" " ind2=" "><subfield code="a">Boca Raton, FL :</subfield><subfield code="b">CRC/Lewis Publishers,</subfield><subfield code="c">1995.</subfield></datafield><datafield tag="300" ind1=" " ind2=" "><subfield code="a">282 pages :</subfield><subfield code="b">illustrations ;</subfield><subfield code="c">24 cm.</subfield></datafield><datafield tag="336" ind1=" " ind2=" "><subfield code="a">text</subfield><subfield code="b">txt</subfield><subfield code="2">rdacontent.</subfield></datafield><datafield tag="337" ind1=" " ind2=" "><subfield code="a">unmediated</subfield><subfield code="b">n</subfield><subfield code="2">rdamedia.</subfield></datafield><datafield tag="338" ind1=" " ind2=" "><subfield code="a">volume</subfield><subfield code="b">nc</subfield><subfield code="2">rdacarrier.</subfield></datafield><datafield tag="504" ind1=" " ind2=" "><subfield code="a">Includes bibliographical references and index.</subfield></datafield><datafield tag="650" ind1=" " ind2="0"><subfield code="a">Pollution.</subfield><subfield code="0">http://id.loc.gov/authorities/subjects/sh85104530.</subfield></datafield><datafield tag="700" ind1="1" ind2=" "><subfield code="a">Moltmann, Johann Felix,</subfield><subfield code="d">1954-</subfield><subfield code="0">http://id.loc.gov/authorities/names/n91002202.</subfield></datafield><datafield tag="991" ind1=" " ind2=" "><subfield code="a">b23774344</subfield></datafield><datafield tag="906" ind1=" " ind2=" "><subfield code="a">0</subfield><subfield code="b">flu</subfield><subfield code="c">multi</subfield><subfield code="d">a</subfield><subfield code="e">r</subfield><subfield code="f">1</subfield><subfield code="g">02-09-96</subfield><subfield code="h">01-10-19</subfield><subfield code="i">20</subfield></datafield><datafield tag="955" ind1=" " ind2=" "><subfield code="u">+l</subfield></datafield><datafield tag="948" ind1=" " ind2=" "><subfield code="a">BSLW 06/11/2018</subfield></datafield><datafield tag="998" ind1=" " ind2=" "><subfield code="a">02/13/96</subfield><subfield code="t">c</subfield><subfield code="s">9110</subfield><subfield code="l">NHDG</subfield><subfield code="n">NhD</subfield><subfield code="w">DCLC9541713B</subfield><subfield code="d">02/13/96</subfield><subfield code="c">JS</subfield><subfield code="b">ACQ</subfield><subfield code="i">960213</subfield></datafield><datafield tag="995" ind1=" " ind2=" "><subfield code="a">OCLC Recl. 2008</subfield></datafield><datafield tag="903" ind1=" " ind2=" "><subfield code="f">corr./JS/2-26-96</subfield></datafield><datafield tag="903" ind1=" " ind2=" "><subfield code="a">QH545.A1</subfield><subfield code="b">R65 1995</subfield></datafield><datafield tag="852" ind1="0" ind2=" "><subfield code="b">DEPOSITORY</subfield><subfield code="c">DEPOSITORY</subfield><subfield code="h">QH545.A1 R65 1996</subfield><subfield code="8">22127176380005706</subfield></datafield><datafield tag="950" ind1=" " ind2=" "><subfield code="z">33311011074376</subfield><subfield code="e">1</subfield></datafield></record>
+</collection>

data/test/tc_controlfield.rb

@@ -1,50 +1,53 @@
-require 'test/unit'
-require 'marc'
+require "test/unit"
+require "marc"
 
 class TestField < Test::Unit::TestCase
-
   def test_control
-    control = MARC::ControlField.new('005', 'foobarbaz')
-    assert_equal(control.to_s, '005 foobarbaz')
+    control = MARC::ControlField.new("005", "foobarbaz")
+    assert_equal(control.to_s, "005 foobarbaz")
   end
 
   def test_field_as_control
-    assert_raise(MARC::Exception) do
-      # can't have a field with a tag < 010
-      field = MARC::DataField.new('007') 
-    end
+    field = MARC::DataField.new("007")
+    assert_equal(field.valid?, false)
   end
 
   def test_alpha_control_field
-    assert_raise(MARC::Exception) do
-      # can't have a field with a tag < 010
-      field = MARC::ControlField.new('DDD') 
-    end
+    # can't have a field with a tag < 010
+    field = MARC::ControlField.new("DDD")
+    assert_equal(field.valid?, false)
   end
-  
+
   def test_extra_control_field
-    MARC::ControlField.control_tags << 'FMT'
-    assert_nothing_raised do
-       field = MARC::ControlField.new('FMT') 
-    end
-    assert_raise(MARC::Exception) do
-      field = MARC::DataField.new('FMT') 
-    end
-    MARC::ControlField.control_tags.delete('FMT')
-    assert_nothing_raised do
-       field = MARC::DataField.new('FMT') 
-    end
-    assert_raise(MARC::Exception) do
-      field = MARC::ControlField.new('FMT') 
-    end
-    
+    MARC::ControlField.control_tags << "FMT"
+    field = MARC::ControlField.new("FMT")
+    assert_equal(field.valid?, true)
+    field = MARC::DataField.new("FMT")
+    assert_equal(field.valid?, false)
+    MARC::ControlField.control_tags.delete("FMT")
+    field = MARC::DataField.new("FMT")
+    assert_equal(field.valid?, true)
+    field = MARC::ControlField.new("FMT")
+    assert_equal(field.valid?, false)
   end
 
   def test_control_as_field
-    assert_raise(MARC::Exception) do
-      # can't have a control with a tag > 009
-      f = MARC::ControlField.new('245')
-    end
+    # can't have a control with a tag > 009
+    f = MARC::ControlField.new("245")
+    assert_equal(f.valid?, false)
   end
-end
 
+  def test_equal
+    f1 = MARC::ControlField.new("001", "foobarbaz")
+    f2 = MARC::ControlField.new("001", "foobarbaz")
+    assert_equal(f1, f2)
+
+    f3 = MARC::ControlField.new("001", "foobarbazqux")
+    assert_not_equal(f1, f3)
+    f4 = MARC::ControlField.new("002", "foobarbaz")
+    assert_not_equal(f1, f4)
+
+    assert_not_equal(f1, "001")
+    assert_not_equal(f2, "foobarbaz")
+  end
+end

data/test/tc_datafield.rb

@@ -1,71 +1,76 @@
-require 'test/unit'
-require 'marc'
+require "test/unit"
+require "marc"
 
 class TestField < Test::Unit::TestCase
+  def test_tag
+    f1 = MARC::DataField.new("100")
+    assert_equal("100", f1.tag)
+    f2 = MARC::DataField.new("100")
+    assert_equal("100", f2.tag)
+    assert_equal(f1, f2)
+    f3 = MARC::DataField.new("245")
+    assert_not_equal(f1, f3)
+  end
 
-    def test_tag
-        f1 = MARC::DataField.new('100')
-        assert_equal('100', f1.tag)
-        f2 = MARC::DataField.new(tag='100')
-        assert_equal('100', f2.tag)
-        assert_equal(f1, f2)
-        f3 = MARC::DataField.new('245')
-        assert_not_equal(f1, f3)
-    end
+  def test_alphabetic_tag
+    alph = MARC::DataField.new("ALF")
+    assert_equal "ALF", alph.tag
 
-    def test_alphabetic_tag
-        alph = MARC::DataField.new('ALF')
-        assert_equal 'ALF', alph.tag
+    alphnum = MARC::DataField.new("0D9")
+    assert_equal "0D9", alphnum.tag
+  end
 
-        alphnum = MARC::DataField.new('0D9')
-        assert_equal '0D9', alphnum.tag
-    end
+  def test_indicators
+    f1 = MARC::DataField.new("100", "0", "1")
+    assert_equal("0", f1.indicator1)
+    assert_equal("1", f1.indicator2)
+    f2 = MARC::DataField.new("100", "0", "1")
+    assert_equal("0", f2.indicator1)
+    assert_equal("1", f2.indicator2)
+    assert_equal(f1, f2)
+    f3 = MARC::DataField.new("100", "1", "1")
+    assert_not_equal(f1, f3)
+  end
 
-    def test_indicators
-        f1 = MARC::DataField.new('100', '0', '1')
-        assert_equal('0', f1.indicator1)
-        assert_equal('1', f1.indicator2)
-        f2 = MARC::DataField.new(tag='100',i1='0',i2='1')
-        assert_equal('0', f2.indicator1)
-        assert_equal('1', f2.indicator2)
-        assert_equal(f1, f2)
-        f3 = MARC::DataField.new(tag='100', i1='1', i2='1')
-        assert_not_equal(f1, f3)
-    end
+  def test_subfields
+    f1 = MARC::DataField.new("100", "0", "1",
+      MARC::Subfield.new("a", "Foo"),
+      MARC::Subfield.new("b", "Bar"))
+    assert_equal("100 01 $a Foo $b Bar ", f1.to_s)
+    assert_equal("FooBar", f1.value)
+    f2 = MARC::DataField.new("100", "0", "1",
+      MARC::Subfield.new("a", "Foo"),
+      MARC::Subfield.new("b", "Bar"))
+    assert_equal(f1, f2)
+    f3 = MARC::DataField.new("100", "0", "1",
+      MARC::Subfield.new("a", "Foo"),
+      MARC::Subfield.new("b", "Bez"))
+    assert_not_equal(f1, f3)
+  end
 
-    def test_subfields
-        f1 = MARC::DataField.new('100', '0', '1', 
-            MARC::Subfield.new('a', 'Foo'),
-            MARC::Subfield.new('b', 'Bar') )
-        assert_equal("100 01 $a Foo $b Bar ", f1.to_s)
-        assert_equal("FooBar", f1.value)
-        f2 = MARC::DataField.new('100', '0', '1', 
-            MARC::Subfield.new('a', 'Foo'),
-            MARC::Subfield.new('b', 'Bar') )
-        assert_equal(f1,f2)
-        f3 = MARC::DataField.new('100', '0', '1', 
-            MARC::Subfield.new('a', 'Foo'),
-            MARC::Subfield.new('b', 'Bez') )
-        assert_not_equal(f1,f3)
-    end
+  def test_subfield_shorthand
+    f = MARC::DataField.new("100", "0", "1", ["a", "Foo"], ["b", "Bar"])
+    assert_equal("100 01 $a Foo $b Bar ", f.to_s)
+  end
 
-    def test_subfield_shorthand
-        f  = MARC::DataField.new('100', '0', '1', ['a', 'Foo'], ['b', 'Bar'])
-        assert_equal('100 01 $a Foo $b Bar ', f.to_s)
-    end
-            
+  def test_iterator
+    field = MARC::DataField.new("100", "0", "1", ["a", "Foo"], ["b", "Bar"],
+      ["a", "Bez"])
+    count = 0
+    field.each { |x| count += 1 }
+    assert_equal(count, 3)
+  end
 
-    def test_iterator
-        field = MARC::DataField.new('100', '0', '1', ['a', 'Foo'],['b', 'Bar'],
-            ['a', 'Bez'])
-        count = 0
-        field.each {|x| count += 1}
-        assert_equal(count,3)
-    end
-    
-    def test_lookup_shorthand
-        f  = MARC::DataField.new('100', '0', '1', ['a', 'Foo'], ['b', 'Bar'])
-        assert_equal(f['b'], 'Bar')
-    end
+  def test_lookup_shorthand
+    f = MARC::DataField.new("100", "0", "1", ["a", "Foo"], ["b", "Bar"])
+    assert_equal(f["b"], "Bar")
+  end
 
+  def test_equal_other_types
+    f = MARC::DataField.new("100", "0", "1",
+      MARC::Subfield.new("a", "Foo"),
+      MARC::Subfield.new("b", "Bar"))
+    assert_not_equal(f, "100 01 $a Foo $b Bar ")
+    assert_not_equal(f, f["a"])
+  end
 end

data/test/tc_dublincore.rb

@@ -1,14 +1,12 @@
-require 'test/unit'
-require 'marc'
+require "test/unit"
+require "marc"
 
 class DublinCoreTest < Test::Unit::TestCase
-
-    def test_mapping
-        reader = MARC::Reader.new('test/batch.dat')
-        reader.each do |record|
-          dc = record.to_dublin_core
-          assert dc['title'] == record['245']['a']
-        end
+  def test_mapping
+    reader = MARC::Reader.new("test/batch.dat")
+    reader.each do |record|
+      dc = record.to_dublin_core
+      assert dc["title"] == record["245"]["a"]
     end
-
-end
+  end
+end

data/test/tc_hash.rb

@@ -1,26 +1,23 @@
-require 'test/unit'
-require 'marc'
-require 'rubygems'
+require "test/unit"
+require "marc"
+require "rubygems"
 
 class TestHash < Test::Unit::TestCase
-
   def test_to_hash
-    raw = IO.read('test/one.dat')
+    raw = IO.read("test/one.dat")
     r = MARC::Record.new_from_marc(raw)
     h = r.to_hash
     assert_kind_of(Hash, h)
-    assert_equal(r.leader, h['leader'])
-    assert_equal(r.fields.length, h['fields'].length)
-    assert_equal(r.fields.first.tag, h['fields'].first.keys.first)
+    assert_equal(r.leader, h["leader"])
+    assert_equal(r.fields.length, h["fields"].length)
+    assert_equal(r.fields.first.tag, h["fields"].first.keys.first)
   end
 
   def test_roundtrip
-    reader = MARC::Reader.new('test/batch.dat')
+    reader = MARC::Reader.new("test/batch.dat")
     reader.each do |r|
       x = MARC::Record.new_from_hash(r.to_hash)
-      assert_equal(r,x)
+      assert_equal(r, x)
     end
   end
-  
-  
-end
+end

data/test/tc_jsonl.rb

@@ -0,0 +1,19 @@
+require "test/unit"
+require "marc"
+
+class JSONLTestCase < Test::Unit::TestCase
+  def test_round_trip
+    tempfile = "test/batch.jsonl"
+    records = MARC::Reader.new("test/batch.dat").to_a
+    jsonl_writer = MARC::JSONLWriter.new(tempfile)
+    records.each { |r| jsonl_writer.write(r) }
+    jsonl_writer.close
+
+    jsonl_reader = MARC::JSONLReader.new(tempfile)
+    jsonl_reader.each_with_index do |r, i|
+      assert_equal(records[i], r)
+    end
+  ensure
+    File.unlink(tempfile)
+  end
+end

data/test/tc_marchash.rb

@@ -1,37 +1,33 @@
-require 'test/unit'
-require 'marc'
-require 'rubygems'
+require "test/unit"
+require "marc"
+require "rubygems"
 
 class TestMARCHASH < Test::Unit::TestCase
-
   def test_simple
     simple = {
-      'type' => 'marc-hash',
-      'version' => [1,0],
-      'leader' => 'LEADER',
-      'fields' => [
-        ['245', '1', '0', 
+      "type" => "marc-hash",
+      "version" => [1, 0],
+      "leader" => "LEADER",
+      "fields" => [
+        ["245", "1", "0",
           [
-            ['a', 'TITLE'],
-            ['b', 'SUBTITLE']
-          ]
-        ]
+            ["a", "TITLE"],
+            ["b", "SUBTITLE"]
+          ]]
       ]
     }
-    r = MARC::Record.new()
-    r.leader = 'LEADER'
-    f = MARC::DataField.new('245', '1', '0', ['a', 'TITLE'], ['b', 'SUBTITLE'])
+    r = MARC::Record.new
+    r.leader = "LEADER"
+    f = MARC::DataField.new("245", "1", "0", ["a", "TITLE"], ["b", "SUBTITLE"])
     r << f
     assert_equal(r.to_marchash, simple)
   end
 
   def test_real
-    reader = MARC::Reader.new('test/batch.dat')
+    reader = MARC::Reader.new("test/batch.dat")
     reader.each do |r|
       x = MARC::Record.new_from_marchash(r.to_marchash)
-      assert_equal(r,x)
+      assert_equal(r, x)
     end
   end
-  
-  
-end
+end