(This is inspired by Python’s xml_objectify)
XMLObject attempts to make the accessing of small, well-formed XML structures convenient, by providing a syntax that fits well in most Ruby programs.
None outside of Ruby, though some optional gems add additional features and speedups. See more below.
I’ve tested this release under:
Rubinius: 1.0.1 (REXML only, libxml-ruby might work if you can build it) MRI: 1.9.2, 1.8.7, 1.8.6 REE: 1.8.7 JRuby: 1.5.2, 1.4.0 (both in 1.8.7 mode)
And with the following optionals:
version 1.1.4 of libxml-ruby (for faster execution) version 2.3.8 of activesupport (for inflected plurals)
gem install xml-object
recipe.xml is as follows:
<recipe name="bread" prep_time="5 mins" cook_time="3 hours">
<title>Basic bread</title>
<ingredient amount="8" unit="dL">Flour</ingredient>
<ingredient amount="10" unit="grams">Yeast</ingredient>
<ingredient amount="4" unit="dL" state="warm">Water</ingredient>
<ingredient amount="1" unit="teaspoon">Salt</ingredient>
<instructions easy="yes" hard="false">
<step>Mix all ingredients together.</step>
<step>Knead thoroughly.</step>
<step>Cover with a cloth, and leave for one hour in warm room.</step>
<step>Knead again.</step>
<step>Place in a bread baking tin.</step>
<step>Cover with a cloth, and leave for one hour in warm room.</step>
<step>Bake in the oven at 180(degrees)C for 30 minutes.</step>
</instructions>
</recipe>
require 'xml-object'
recipe = XMLObject.new(File.open('recipe.xml'))
recipe.name => "bread"
recipe.title => "Basic bread"
recipe.ingredients.is_a?(Array) => true
recipe.ingredients.first.amount => "8" # Not a Fixnum. Too hard. :(
recipe.instructions.easy? => true
recipe.instructions.first.upcase => "MIX ALL INGREDIENTS TOGETHER."
recipe.instructions.steps.size => 7
XML is an extensible markup language. It is extensible because it is meant to define markup languages for any type of document, so new tags are needed depending on the problem domain.
Sometimes, however, XML ends up being used to solve a much simpler problem: the issue of passing a data-structure over the network, and/or between two different languages. Tools like JSON or YAML are a much better fit for this kind of job, but one doesn’t always have that luxury.
XMLObject supports different adapters to do the actual XML parsing. It ships with REXML, and LibXML adapters. By default, the REXML adapter is used.
To use a different adapter than the REXML default:
require 'xml-object' # Require XMLObject first require 'xml-object/adapters/libxml'
XMLObject uses dot notation (foo.bar) for both elements and attributes, with a few rules, and with an array notation fallback for invalid method names or other tricky situations. For example, with the given file:
<outer object_id="root" name="foo"> <name>Outer Element</name> </outer>
outer.name is the name element. Child elements are always looked up first, then attributes. To access the attribute in case of ambiguity, use outer[:attr => ‘name’].
outer.object_id is really Object#object_id, because all of the object methods are preserved (this is on purpose). To access the attribute object_id, use outer[:attr => ‘object_id’] (or just outer, since there’s no element/attribute ambiguity there).
Elements or attributes that look like booleans are “booleanized” if called by their question names (such as enabled?)
Like XmlSimple, XMLObject folds same-named elements found at the same level, like so:
<student> <name>Bob</name> <course>Math</course> | <course>Biology</course> | => 'course' becomes an Array </student> student = XMLObject.new(xml_file) student.course.is_a?(Array) => true student.course.first == 'Math' => true student.course.last == 'Biology => true
With the same file as in the example above:
student.courses.first == student.course.first => true
Note that the pluralization algorithm is just tacking an ‘s’ at the end of the singular, unless ActiveSupport is required, in which case you get inflected plurals.
Sometimes, collections are expressed with a container element in XML:
<author>
<name>John</name>
<publications>
<book>Math 101</book>
<book>Biology 101</book>
</publications>
</author>
In this case, since the container element courses has no text element of its own (attributes are ok), and it only has elements of one name under it, it delegates all methods to the collection below, so you get:
author.publications == author.publications.books => true author.publications.map { |b| b.downcase } => ['math 101', 'biology 101']
The design of the adapters assumes parsing of the objects recursively. Deep files are bound to throw SystemStackError, but for the kinds of files I need to read, things are working fine so far. In any case, stream parsing is on the TODO list.
It most likely doesn’t work with a ton of features of complex XML files (see the caveats section). I’ll always try to accommodate those, as long as they don’t make the basic usage more complex. As usual, patches welcome.
The LibXML adapter will not return the ‘xmlns’ attribute.
Copyright © 2008, 2009 Jordi Bunster, released under the MIT license