Empact / roxml

ROXML is a module for binding Ruby classes to XML. It supports custom mapping and bidirectional marshalling between Ruby and XML using annotation-style class methods, via Nokogiri or LibXML.

This URL has Read+Write access

roxml /
README.rdoc

ROXML Ruby Object to XML mapping library. For more information visit roxml.rubyforge.org

Quick Start Guide

This is a short usage example. See ROXML::ClassMethods::Declarations and packaged test cases for more information.

Basic Mapping

Consider an XML document representing a Library containing a number of Books. You can map this structure to Ruby classes that provide addition useful behavior. With ROXML, you can annotate the Ruby classes as follows:

  class Book
    include ROXML

    xml_accessor :isbn, :from => "@ISBN" # attribute with name 'ISBN'
    xml_accessor :title
    xml_accessor :description, :cdata => true  # text node with cdata protection
    xml_accessor :author
  end

  class Library
    include ROXML

    xml_accessor :name, :from => "NAME", :cdata => true
    xml_accessor :books, :as => [Book] # by default roxml searches for books for in <book> child nodes, then, if none are present, in ./books/book children
  end

To create a library and put a number of books in it we could run the following code:

  book = Book.new
  book.isbn = "0201710897"
  book.title = "The PickAxe"
  book.description = "Best Ruby book out there!"
  book.author = "David Thomas, Andrew Hunt, Dave Thomas"

  lib = Library.new
  lib.name = "Favorite Books"
  lib.books = [book]

To save this information to an XML file:

  doc = ROXML::XML::Document.new
  doc.root = lib.to_xml
  doc.save("library.xml")

To later populate the library object from the XML file:

  lib = Library.from_xml(File.read("library.xml"))

Similarly, to do a one-to-one mapping between XML objects, such as book and publisher, you would add a reference to another ROXML class. For example:

  <book isbn="0974514055">
    <title>Programming Ruby - 2nd Edition</title>
    <description>Second edition of the great book.</description>
    <publisher>
      <name>Pragmatic Bookshelf</name>
    </publisher>
  </book>

can be mapped using the following code:

  class Publisher
    include ROXML

    xml_accessor :name

    # other important functionality
  end

  class BookWithPublisher
    include ROXML

    xml_name 'book'
    xml_reader :publisher, :as => Publisher

    #  or, alternatively, if no class is needed to hang functionality on:
    # xml_reader :publisher, :from => 'name', :in => 'publisher'
  end

Note: In the above example, xml_name annotation tells ROXML to set the element name to "book" for mapping to XML. The default is XML element name is the class name in lowercase; "bookwithpublisher" in this case.

Manipulation

Extending the above examples, say you want to parse a book’s page count and have it available as an Integer. In such a case, you can extend any object with a block to manipulate it’s value at parse time. For example:

  class Dog
    include ROXML

    xml_reader(:age, :from => '@human_years', :as => Integer) {|years| years * 7 }
  end

The result of the block above is stored, rather than the actual value parsed from the document.

Construction

Object life-cycle is as follows: .from_xml is called with a first argument representing the xml in file, string, or path form, and with optional initialization_args following.

Firt .new and thus #initialize, is called with those same initialization_args, or no args if none are present. Then the object is populated with the attribute values from xml. Then the #after_parse callback is called, with no arguments.

In #after_parse you can ensure that your object initialization is complete, including initialization which requires more than one variable in concert.

E.g.:

  class Measurement
    include ROXML

    xml_reader :units, :from => :attr
    xml_reader :value, :from => :content

    def initialize(value = 0, units = 'meters')
      to_metric
    end

  private
    def after_parse
      # xml attributes of self are already valid
      to_metric
    end

    def to_metric
      # translate units & value into metric, for example
    end
  end

One important use of this approach is to make ROXML object which may or may not include an xml backing, which may be used via new construction as well as from_xml construction.

Selecting a parser

By default, ROXML will use LibXML if it is available, or otherwise REXML. If you’d like to explicitly require one or the other, you may do the following:

  module ROXML
    XML_PARSER = 'libxml' # or 'rexml'
  end
  require 'roxml'

For more information on available annotations, see ROXML::ClassMethods::Declarations