marc 0.1.5 → 0.1.6

data/lib/marc/datafield.rb

@@ -6,12 +6,13 @@ module MARC
   # MARC records contain data fields, each of which has a tag, 
   # indicators and subfields. Tags for data fields must be in
   # the range 010-999.
-  # Accessor attributes: tag ; indicator1 ; indicator2
+  # Accessor attributes: tag, indicator1, indicator2
   # 
-  # DataField includes enumerable for access to it's constituent
+  # DataField mixes in Enumerable to enable access to it's constituent
   # Subfield objects. For instance, if you have a DataField representing
   # a 856 tag, and want to find all 'z' subfields:
-  # urltag.find_all { |subfield| subfield.code == 'z' }
+  #
+  #   subfield_z = field.find_all {|subfield| subfield.code == 'z'}
   #
   # Also, the accessor 'subfields' is an array of MARC::Subfield objects
   # which can be accessed or modified by the client directly if

data/lib/marc/reader.rb

@@ -37,11 +37,17 @@ module MARC
 
     def each 
       # while there is data left in the file
-      while length = @handle.read(5)
+      while rec_length_s = @handle.read(5)
+        # make sure the record length looks like an integer
+        rec_length_i = rec_length_s.to_i
+        if rec_length_i == 0:
+          raise MARC::Exception.new("invalid record length: #{rec_length_s}")
+        end
 
         # get the raw MARC21 for a record back from the file
         # using the record length
-        raw = length + @handle.read(length.to_i-5)
+        raw = rec_length_s + @handle.read(rec_length_i-5)
+        
 
         # create a record from the data and return it
         #record = MARC::Record.new_from_marc(raw)

data/lib/marc/record.rb

@@ -3,12 +3,15 @@ module MARC
   # A class that represents an individual MARC record. Every record
   # is made up of a collection of MARC::DataField objects. 
   #
-  # MARC::Record includes Enumerable for access to constituent
-  # DataFields. Eg, to return a list of all 650 DataFields:
-  # record.find_all { |field| field.tag == '650' }  
+  # MARC::Record mixes in Enumerable to enable access to constituent
+  # DataFields. For example, to return a list of all subject DataFields:
+  #
+  #   record.find_all {|field| field.tag =~ /^6../}  
   # 
   # The accessor 'fields' is also an Array of MARC::DataField objects which
   # the client can access or modifyi if neccesary.
+  #
+  #   record.fields.delete(field)
   # 
   # Other accessor attribute: 'leader' for record leader as String
  
@@ -16,10 +19,10 @@ module MARC
     include Enumerable
 
     # the record fields
-    attr_accessor :fields,
+    attr_accessor :fields
 
     # the record leader
-    :leader
+    attr_accessor :leader
 
     def initialize
       @fields = []
@@ -38,6 +41,12 @@ module MARC
       @fields.push(field)
     end
 
+    # alias to append
+    
+    def <<(field)
+      append(field)
+    end
+
     # each() is here to support iterating and searching since MARC::Record
     # mixes in Enumberable
     #

data/lib/marc/subfield.rb

@@ -1,9 +1,9 @@
 module MARC
   
-  # A class to represents a subfield within a DataField. 
-  # Accessor attributes: code (letter subfield code) ; value
-  # Both can be empty string, but should not be nil. 
-
+  # A class that represents an individual  subfield within a DataField. 
+  # Accessor attributes include: code (letter subfield code) and value 
+  # (the content of the subfield). Both can be empty string, but should 
+  # not be set to nil. 
 
   class Subfield
     attr_accessor :code, :value

data/lib/marc/xmlreader.rb

@@ -60,7 +60,7 @@ module MARC
         event = @parser.pull
 
         if event.text?
-          text += event[0].strip
+          text += REXML::Text::unnormalize(event[0].strip)
           next
         end
 

data/lib/marc/xmlwriter.rb

@@ -4,6 +4,8 @@ require 'rexml/text'
 module MARC
   
   # A class for writing MARC records as MARCXML.
+  # BIG CAVEAT! XMLWriter will *not* convert your MARC8 to UTF8
+  # bug the authors to do this if you need it
   
   class XMLWriter
     
@@ -51,21 +53,6 @@ module MARC
       @fh.close
     end
 
-
-    # Converts from ISO 8859-1 to UTF-8, normalizes the UTF-8, and puts a
-    # 'clean up marker' in records that have control characters (which are
-    # not valid in XML). This is useful for locating these records once
-    # they are in XML so problems caused by removing the invalid characters
-    # can be fixed by a person.  This (or something in the module) needs to
-    # convert from MARC-8 to UTF-8, but it doesn't do this yet...
-
-    def self.convert_to_utf8(text)
-      cleaned_text = text.gsub(/[\x00-\x1f\x7f-\xff]+/, ' CLEAN_ME_UP ')
-      utf8_text = cleaned_text.unpack('C*').pack('U*')
-      normalized_text = REXML::Text::normalize(utf8_text)
-
-      return normalized_text
-    end
     
     # a static method that accepts a MARC::Record object
     # and returns a REXML::Document for the XML serialization.
@@ -138,7 +125,7 @@ module MARC
             end
             
             subfield_element.add_attribute("code", subfield.code)
-            text = MARC::XMLWriter.convert_to_utf8(subfield.value)
+            text = subfield.value
             subfield_element.add_text(text)
             datafield_elem.add_element(subfield_element)
           end
@@ -153,7 +140,7 @@ module MARC
           end
           
           control_element.add_attribute("tag", field.tag)
-          text = MARC::XMLWriter.convert_to_utf8(field.value)
+          text = field.value
           control_element.add_text(text)
           e.add_element(control_element)
         end

data/test/tc_reader.rb

@@ -3,36 +3,41 @@ require 'marc'
 
 class ReaderTest < Test::Unit::TestCase
 
-    def test_batch
-        reader = MARC::Reader.new('test/batch.dat')
-        count = 0
-        reader.each { count += 1 }
-        assert_equal(count, 10)
-    end
-
-    def test_loose
-        reader = MARC::ForgivingReader.new('test/batch.dat')
-        count = 0
-        reader.each { count += 1 }
-        assert_equal(10, count)
-    end
-
-    def test_search
-        reader = MARC::Reader.new('test/batch.dat')
-        records = reader.find_all { |r| r =~ /Perl/ }
-        assert_equal(10, records.length)
-
-        reader = MARC::Reader.new('test/batch.dat')
-        records = reader.find_all { |r| r['245'] =~ /Perl/ }
-        assert_equal(10, records.length)
-
-        reader = MARC::Reader.new('test/batch.dat')
-        records = reader.find_all { |r| r['245']['a'] =~ /Perl/ }
-        assert_equal(10, records.length)
-
-        reader = MARC::Reader.new('test/batch.dat')
-        records = reader.find_all { |r| r =~ /Foo/ }
-        assert_equal(0, records.length)
-    end
+  def test_batch
+    reader = MARC::Reader.new('test/batch.dat')
+    count = 0
+    reader.each { count += 1 }
+    assert_equal(count, 10)
+  end
+
+  def test_loose
+    reader = MARC::ForgivingReader.new('test/batch.dat')
+    count = 0
+    reader.each { count += 1 }
+    assert_equal(10, count)
+  end
+
+  def test_bad_marc
+    reader = MARC::Reader.new('test/tc_reader.rb')
+    assert_raises(MARC::Exception) {reader.entries[0]}
+  end
+
+  def test_search
+    reader = MARC::Reader.new('test/batch.dat')
+    records = reader.find_all { |r| r =~ /Perl/ }
+    assert_equal(10, records.length)
+
+    reader = MARC::Reader.new('test/batch.dat')
+    records = reader.find_all { |r| r['245'] =~ /Perl/ }
+    assert_equal(10, records.length)
+
+    reader = MARC::Reader.new('test/batch.dat')
+    records = reader.find_all { |r| r['245']['a'] =~ /Perl/ }
+    assert_equal(10, records.length)
+
+    reader = MARC::Reader.new('test/batch.dat')
+    records = reader.find_all { |r| r =~ /Foo/ }
+    assert_equal(0, records.length)
+  end
 
 end

data/test/tc_xml.rb

@@ -4,6 +4,17 @@ require 'stringio'
 
 class XMLTest < Test::Unit::TestCase
 
+  def test_xml_entities
+    r1 = MARC::Record.new
+    r1 << MARC::DataField.new('245', '0', '0', ['a', 'foo & bar'])
+    xml = r1.to_xml.to_s
+    assert_match /foo &amp; bar/, xml
+
+    reader = MARC::XMLReader.new(StringIO.new(xml))
+    r2 = reader.entries[0]
+    assert_equal 'foo & bar', r2['245']['a']
+  end
+
   def test_batch
     reader = MARC::XMLReader.new('test/batch.xml')
     count = 0

metadata

@@ -3,8 +3,8 @@ rubygems_version: 0.9.2
 specification_version: 1
 name: marc
 version: !ruby/object:Gem::Version 
-  version: 0.1.5
-date: 2007-05-01 00:00:00 -04:00
+  version: 0.1.6
+date: 2007-05-04 00:00:00 -04:00
 summary: A ruby library for working with Machine Readable Cataloging
 require_paths: 
 - lib