rbook-onix 0.5.2 → 0.6.0

data/lib/rbook/onix/sales_restriction.rb

@@ -6,50 +6,45 @@ module Onix
   class SalesRestriction
     attr_accessor :type, :detail
 
-    # Attempts to create a contributor object using the supplied xml element
-    def SalesRestriction.load_from_element(element)
-      raise ArgumentError, 'load_from_element expects a REXML element object' unless element.class == REXML::Element
-      
-      if REXML::XPath.first(element, '//SalesRestriction').nil? 
-        raise LoadError, 'supplied REXML document object does not appear contain a valid SalesRestriction fragment'
-      end
-      
+    def self.load_from_string(str)
+      doc = Hpricot::XML(str)
       restriction = SalesRestriction.new
 
-      tmp = REXML::XPath.first(element, '//SalesRestriction/SalesRestrictionType')
-      restriction.type = tmp.text if tmp != nil
+      tmp = doc.search('//SalesRestriction/SalesRestrictionType')
+      restriction.type = tmp.inner_html unless tmp.inner_html.blank?
 
-      tmp = REXML::XPath.first(element, '//SalesRestriction/SalesRestrictionDetail')
-      restriction.detail = tmp.text if tmp != nil
+      tmp = doc.search('//SalesRestriction/SalesRestrictionDetail')
+      restriction.detail = tmp.inner_html unless tmp.inner_html.blank?
 
       return restriction
-
     end
- 
-    # Return an xml element representing this contributor
-    def to_element
-      raise 'SalesRestriction must have a type to create an element' if self.type.nil?
-      raise 'Contributor must have a detail to create an element' if self.detail.nil?
+
+    # Attempts to create a contributor object using the supplied xml element
+    def SalesRestriction.load_from_element(element)
+      raise ArgumentError, 'load_from_element expects a REXML element object' unless element.class == REXML::Element
       
-      restriction =  REXML::Element.new('SalesRestriction') 
-      restriction.add_element('SalesRestrictionType').text = self.type.to_xs
-      restriction.add_element('SalesRestrictionDetail').text = self.detail.to_xs
+      if REXML::XPath.first(element, '//SalesRestriction').nil? 
+        raise LoadError, 'supplied REXML document object does not appear contain a valid SalesRestriction fragment'
+      end
       
-      return restriction
+      self.load_from_string(element.to_s)
     end
  
     # Return an XML string representing this contributor
     def to_s
-      output = ''
-
-      # handle the REXML API change in 1.8.6.110. bah!
-      if RUBY_VERSION >= "1.8.6" && RUBY_PATCHLEVEL >= 110
-        formatter = REXML::Formatters::Pretty.new( 2 )
-        formatter.write( to_element , output )
-      else
-        to_element.write(output, 0)
+      raise 'SalesRestriction must have a type to create an element' if self.type.nil?
+      raise 'Contributor must have a detail to create an element' if self.detail.nil?
+
+      builder = Builder::XmlMarkup.new(:indent => 2)
+
+      builder.SalesRestriction do |sr|
+        sr.SalesRestrictionType self.type
+        sr.SalesRestrictionDetail self.detail
       end
-      return output
+    end
+
+    def to_element
+      REXML::Document.new(to_s).root
     end
   end
 end

data/lib/rbook/onix/stream_reader.rb

@@ -1,5 +1,6 @@
 require 'rexml/document'
 require 'thread'
+require 'timeout'
 
 module RBook
   module Onix
@@ -14,6 +15,11 @@ module RBook
       end
 
       def doctype(name, pub_sys, long_name, uri)   
+        # do nothing
+      end
+
+      def instruction(name, instruction)
+        # do nothing
       end
 
       def tag_start(name, attrs)
@@ -29,7 +35,7 @@ module RBook
       end
 
       def text(text)
-        @product_fragment << text if @in_product
+        @product_fragment << text.to_xs if @in_product
       end
 
       def tag_end(name)
@@ -43,13 +49,13 @@ module RBook
           # A product tag is finished, so add it to the queue
           @product_fragment << "</Product>"
           begin
-            doc = REXML::Document.new(@product_fragment)
-            unless doc.root.nil?
-              product = RBook::Onix::Product.load_from_element(doc.root)
-              @queue.push(product) unless product.nil?
-            end
-          rescue
+            product = RBook::Onix::Product.load_from_string(@product_fragment)
+            @queue.push(product) unless product.nil?
+          rescue Exception => e
             # error occurred while building the product from an XML fragment
+            # pop the error on the queue so it can be raised by the thread
+            # reading items off the queue
+            @queue.push(e)
           end
           @in_product = false
         else
@@ -61,7 +67,14 @@ module RBook
         # do nothing
       end
 
-      def method_missing
+      def end_document
+        # signal to the thread reading products off the queue that there are none left.
+        # this event curently isn't supported by rexml, so if an ONIX file is truncated
+        # and is missing </ONIXMessage>, it will hang. Patch submitted upstream
+        @queue.push(nil)
+      end
+
+      def method_missing(*args)
         # do nothing
       end
     end
@@ -91,6 +104,7 @@ module RBook
         else
           throw "Unable to read from path or file"                
         end
+
         # create a sized queue to store each product read from the file
         @queue = SizedQueue.new(100)
 
@@ -109,11 +123,19 @@ module RBook
       #    puts product.inspect
       #  end
       def each
-        obj = @queue.pop
+        # if the ONIX file we're processing has been truncated (no </ONIXMessage>), then 
+        # we will block on the next @queue.pop indefinitely, so give it a time limit
+        obj = nil
+        Timeout::timeout(5) { obj = @queue.pop }
         while !obj.nil?
+          raise obj if obj.kind_of?(Exception)
           yield obj
-          obj = @queue.pop
+
+          Timeout::timeout(5) { obj = @queue.pop }
         end
+      rescue Timeout::Error
+        # do nothing, no more items on the queue - possibly the source
+        # file wasn't an XML file?
       end
     end
   end  

data/lib/rbook/onix/supply_detail.rb

@@ -7,74 +7,70 @@ module Onix
   class SupplyDetail
     attr_accessor :supplier_name, :availability_code, :intermediary_availability_code, :price, :price_type_code
 
-    # Attempts to create a contributor object using the supplied xml element
-    def self.load_from_element(element)
-      raise ArgumentError, 'load_from_element expects a REXML element object' unless element.class == REXML::Element
-      
-      if REXML::XPath.first(element, '//SupplyDetail').nil? 
-        raise LoadError, 'supplied REXML document object does not appear contain a valid SupplyDetail fragment'
-      end
-      
+    def self.load_from_string(str)
+      doc = Hpricot::XML(str)
       supply_detail = SupplyDetail.new
 
-      tmp = REXML::XPath.first(element, '//SupplyDetail/SupplierName')
-      supply_detail.supplier_name = tmp.text if tmp != nil
+      tmp = doc.search('//SupplyDetail/SupplierName')
+      supply_detail.supplier_name = tmp.inner_html unless tmp.inner_html.blank?
 
-      tmp = REXML::XPath.first(element, '//SupplyDetail/AvailabilityCode')
-      supply_detail.availability_code = tmp.text if tmp != nil
+      tmp = doc.search('//SupplyDetail/AvailabilityCode')
+      supply_detail.availability_code = tmp.inner_html unless tmp.inner_html.blank?
 
-      tmp = REXML::XPath.first(element, '//SupplyDetail/IntermediaryAvailabilityCode')
-      supply_detail.intermediary_availability_code = tmp.text if tmp != nil
+      tmp = doc.search('//SupplyDetail/IntermediaryAvailabilityCode')
+      supply_detail.intermediary_availability_code = tmp.inner_html unless tmp.inner_html.blank?
       
-      tmp = REXML::XPath.first(element, '//SupplyDetail/Price/PriceAmount')
-      supply_detail.price = BigDecimal(tmp.text) if tmp != nil
+      tmp = doc.search('//SupplyDetail/Price/PriceAmount')
+      supply_detail.price = BigDecimal(tmp.inner_html) unless tmp.inner_html.blank?
 
-      tmp = REXML::XPath.first(element, '//SupplyDetail/Price/PriceTypeCode')
-      supply_detail.price_type_code = tmp.text.to_i if tmp != nil
+      tmp = doc.search('//SupplyDetail/Price/PriceTypeCode')
+      supply_detail.price_type_code = tmp.inner_html.to_i unless tmp.inner_html.blank?
 
       return supply_detail
-
     end
- 
-    # Return an xml element representing this contributor
-    def to_element
-      raise 'SupplyDetail must have a supplier name to create an element' if self.supplier_name.nil?
-      raise 'SupplyDetail must have a availability code to create an element' if self.availability_code.nil?
+
+    # Attempts to create a contributor object using the supplied xml element
+    def self.load_from_element(element)
+      raise ArgumentError, 'load_from_element expects a REXML element object' unless element.class == REXML::Element
       
-      supply_detail =  REXML::Element.new('SupplyDetail') 
-      supply_detail.add_element('SupplierName').text = self.supplier_name.to_xs
-      supply_detail.add_element('AvailabilityCode').text = self.availability_code.to_xs unless self.availability_code.nil?
-      supply_detail.add_element('IntermediaryAvailabilityCode').text = self.intermediary_availability_code.to_xs unless self.intermediary_availability_code.nil?
-      unless self.price.nil?
-        tmp = REXML::Element.new('Price')
-        if self.price.kind_of?(BigDecimal)
-          tmp.add_element('PriceAmount').text = self.price.to_s("F")
-        else
-          tmp.add_element('PriceAmount').text = self.price.to_s
-        end
-        supply_detail.add_element(tmp)
-        if self.price_type_code.nil?
-          tmp.add_element('PriceTypeCode').text = "02"
-        else
-          tmp.add_element('PriceTypeCode').text = self.price_type_code
-        end
+      if REXML::XPath.first(element, '//SupplyDetail').nil? 
+        raise LoadError, 'supplied REXML document object does not appear contain a valid SupplyDetail fragment'
       end
       
-      return supply_detail
+      self.load_from_string(element.to_s)
     end
  
-    # Return an XML string representing this contributor
+    def to_element
+      REXML::Document.new(to_s).root
+    end
+
+    # Return an XML string representing this supply detail
     def to_s
-      output = ''
+      raise 'SupplyDetail must have a supplier name to create an element' if self.supplier_name.nil?
+      raise 'SupplyDetail must have a availability code to create an element' if self.availability_code.nil?
+      
+      builder = Builder::XmlMarkup.new(:indent => 2)
 
-      # handle the REXML API change in 1.8.6.110. bah!
-      if RUBY_VERSION >= "1.8.6" && RUBY_PATCHLEVEL >= 110
-        formatter = REXML::Formatters::Pretty.new( 2 )
-        formatter.write( to_element , output )
-      else
-        to_element.write(output, 0)
+      builder.SupplyDetail do |supply|
+        supply.SupplierName self.supplier_name
+        supply.AvailabilityCode self.availability_code unless self.availability_code.nil?
+        supply.IntermediaryAvailabilityCode self.intermediary_availability_code unless self.intermediary_availability_code.nil?
+        unless self.price.nil?
+          tmp = REXML::Element.new('Price')
+          supply.Price do |p|
+            if self.price.kind_of?(BigDecimal)
+              p.PriceAmount self.price.to_s("F")
+            else
+              p.PriceAmount self.price.to_s
+            end
+            if self.price_type_code.nil?
+              p.PriceTypeCode "02"
+            else
+              p.PriceTypeCode self.price_type_code
+            end
+          end
+        end
       end
-      return output
     end
   end
   end

data/specs/contributor_with_data_spec.rb

@@ -16,12 +16,12 @@ context "A product object with valid data" do
     contributor.sequence_number = '01'
     contributor.role = 'A01'
 
-    doc = contributor.to_element
+    doc = REXML::Document.new(contributor.to_s)
     
-    REXML::XPath.first(doc, '/PersonNameInverted').should_not be_nil
-    REXML::XPath.first(doc, '/PersonNameInverted').text.should eql("Healy, James")
-    REXML::XPath.first(doc, '/ContributorRole').should_not be_nil
-    REXML::XPath.first(doc, '/ContributorRole').text.should eql("A01")
+    REXML::XPath.first(doc, '//PersonNameInverted').should_not be_nil
+    REXML::XPath.first(doc, '//PersonNameInverted').text.should eql("Healy, James")
+    REXML::XPath.first(doc, '//ContributorRole').should_not be_nil
+    REXML::XPath.first(doc, '//ContributorRole').text.should eql("A01")
   end
 
   specify "should output a valid string" do
@@ -31,6 +31,14 @@ context "A product object with valid data" do
     contributor.role = 'A01'
     
     contributor.to_s.should be_a_kind_of(String)
+  end
 
+  specify "should output a valid REXML::Element" do
+    contributor = RBook::Onix::Contributor.new
+    contributor.name_inverted = 'Healy, James'
+    contributor.sequence_number = '01'
+    contributor.role = 'A01'
+    
+    contributor.to_element.should be_a_kind_of(REXML::Element)
   end
 end

data/specs/data/2_ids.xml

@@ -0,0 +1,55 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!DOCTYPE ONIXMessage SYSTEM "http://www.editeur.org/onix/2.1/02/reference/onix-international.dtd">
+<ONIXMessage>
+<Header>
+<FromCompany>Walker Books Australia</FromCompany>
+<SentDate>20070613</SentDate>
+</Header>
+<Product>
+<RecordReference>9789900036708</RecordReference>
+<NotificationType>03</NotificationType>
+<ProductIdentifier>
+<ProductIDType>03</ProductIDType>
+<IDValue>9789900036708</IDValue>
+</ProductIdentifier>
+<ProductIdentifier>
+<ProductIDType>15</ProductIDType>
+<IDValue>9789900036708</IDValue>
+</ProductIdentifier>
+<ProductForm>WX</ProductForm>
+<NoContributor/>
+<AudienceCode>02</AudienceCode>
+<Imprint>
+<ImprintName>Walker Books</ImprintName>
+</Imprint>
+<Publisher>
+<PublishingRole>01</PublishingRole>
+<PublisherName>Walker Books</PublisherName>
+</Publisher>
+<PublishingStatus>02</PublishingStatus>
+<PublicationDate>20070901</PublicationDate>
+<SupplyDetail>
+<SupplierName>TL Distribution</SupplierName>
+<SupplierRole>02</SupplierRole>
+<ProductAvailability>10</ProductAvailability>
+<Stock>
+<OnHand>NYP</OnHand>
+<OnOrder>No</OnOrder>
+</Stock>
+<Price>
+<PriceTypeCode>02</PriceTypeCode>
+<PriceAmount>508.50</PriceAmount>
+</Price>
+</SupplyDetail>
+<MarketRepresentation>
+<AgentName>Walker Books Australia</AgentName>
+<AgentRole>07</AgentRole>
+<MarketCountry>AU</MarketCountry>
+<MarketPublishingStatus>02</MarketPublishingStatus>
+<MarketDate>
+<MarketDateRole>01</MarketDateRole>
+<Date>20070901</Date>
+</MarketDate>
+</MarketRepresentation>
+</Product>
+</ONIXMessage>

data/specs/data/leading_garbage.xml

@@ -0,0 +1,70 @@
+<b>ERROR!!!!</b> ORA-00942: table or view does not exist
+<br><b>ERROR!!!!</b> ORA-00942: table or view does not exist
+<br><b>ERROR!!!!</b> ORA-00942: table or view does not exist
+<br><b>ERROR!!!!</b> ORA-00942: table or view does not exist
+<br><b>ERROR!!!!</b> ORA-00942: table or view does not exist
+<br><?xml version="1.0" encoding="utf-8"?>
+<!DOCTYPE ONIXMessage SYSTEM "http://www.editeur.org/onix/2.1/reference/onix-international.dtd">
+<ONIXMessage>
+  <Header>
+    <FromCompany>TitlePage</FromCompany>
+    <FromPerson>TitlePage 02 92819788</FromPerson>
+    <FromEmail>titlepage@publishers.asn.au</FromEmail>
+    <SentDate>20071104</SentDate>
+    <MessageNote>This data is copyright to TitlePage. TitlePage makes no guarantee of the accuracy or the timeliness of production. It is supplied for your exclusive use as our customer and is only to be used for advising on the pricing and availability of publications (Authorised Purpose). You must not charge a fee for this service. TitlePage and its content may not be used for any other purpose. While it may be copied once for the authorised purpose, written permission from TitlePage must be obtained for any other use. If you were not an intended recipient, you must notify the sender and delete all copies.</MessageNote>
+  </Header>
+  <Product>
+    <RecordReference>77-9780140441185</RecordReference>
+    <NotificationType>03</NotificationType>
+    <ProductForm>BC</ProductForm>
+    <ProductFormDetail>B305</ProductFormDetail>
+    <Title>
+      <TitleType>01</TitleType>
+      <TitleText>Thus Spoke Zarathustra</TitleText>
+    </Title>
+    <Website>
+      <WebsiteLink>http://www.penguin.com.au/default.cfm?SBN=9780140441185</WebsiteLink>
+    </Website>
+    <EditionNumber>1</EditionNumber>
+    <NumberOfPages>352</NumberOfPages>
+    <BICMainSubject>HPC</BICMainSubject>
+    <AudienceCode>01</AudienceCode>
+    <MediaFile>
+      <MediaFileTypeCode>04</MediaFileTypeCode>
+      <MediaFileLinkTypeCode>01</MediaFileLinkTypeCode>
+      <MediaFileLink>http://coverimages.titlepage.com.au/77/9780140441185.gif</MediaFileLink>
+    </MediaFile>
+    <Imprint>
+      <ImprintName>Penguin</ImprintName>
+    </Imprint>
+    <Publisher>
+      <PublishingRole>01</PublishingRole>
+      <PublisherName>Penguin UK</PublisherName>
+    </Publisher>
+    <PublishingStatus>04</PublishingStatus>
+    <PublicationDate>19640101</PublicationDate>
+    <YearFirstPublished>1964</YearFirstPublished>
+    <SupplyDetail>
+      <SupplierName>United Book Distributors</SupplierName>
+      <ProductAvailability>20</ProductAvailability>
+      <Stock>
+        <OnHand>In Stock</OnHand>
+        <OnOrder>No</OnOrder>
+      </Stock>
+      <PackQuantity>48</PackQuantity>
+      <Price>
+        <PriceTypeCode>02</PriceTypeCode>
+        <PriceAmount>12.95</PriceAmount>
+      </Price>
+    </SupplyDetail>
+    <MarketRepresentation>
+      <AgentName>Penguin Group Australia</AgentName>
+      <MarketCountry>AU</MarketCountry>
+      <MarketPublishingStatus>04</MarketPublishingStatus>
+      <MarketDate>
+        <MarketDateRole>01</MarketDateRole>
+        <Date>19640101</Date>
+      </MarketDate>
+    </MarketRepresentation>
+  </Product>
+</ONIXMessage>