Nokogiri 鋸

Nokogiri builder can be used for building XML and HTML documents.

Synopsis:

builder = Nokogiri::XML::Builder.new do |xml|
  xml.root {
    xml.products {
      xml.widget {
        xml.id_ "10"
        xml.name "Awesome widget"
      }
    }
  }
end
puts builder.to_xml

Will output:

<?xml version="1.0"?>
<root>
  <products>
    <widget>
      <id>10</id>
      <name>Awesome widget</name>
    </widget>
  </products>
</root>

Builder scope

The builder allows two forms. When the builder is supplied with a block that has a parameter, the outside scope is maintained. This means you can access variables that are outside your builder. If you don’t need outside scope, you can use the builder without the “xml” prefix like this:

builder = Nokogiri::XML::Builder.new do
  root {
    products {
      widget {
        id_ "10"
        name "Awesome widget"
      }
    }
  }
end

Special Tags

The builder works by taking advantage of method_missing. Unfortunately some methods are defined in ruby that are difficult or dangerous to remove. You may want to create tags with the name “type”, “class”, and “id” for example. In that case, you can use an underscore to disambiguate your tag name from the method call.

Here is an example of using the underscore to disambiguate tag names from ruby methods:

@objects = [Object.new, Object.new, Object.new]

builder = Nokogiri::XML::Builder.new do |xml|
  xml.root {
    xml.objects {
      @objects.each do |o|
        xml.object {
          xml.type_   o.type
          xml.class_  o.class.name
          xml.id_     o.id
        }
      end
    }
  }
end
puts builder.to_xml

The underscore may be used with any tag name, and the last underscore will just be removed. This code will output the following XML:

<?xml version="1.0"?>
<root>
  <objects>
    <object>
      <type>Object</type>
      <class>Object</class>
      <id>48390</id>
    </object>
    <object>
      <type>Object</type>
      <class>Object</class>
      <id>48380</id>
    </object>
    <object>
      <type>Object</type>
      <class>Object</class>
      <id>48370</id>
    </object>
  </objects>
</root>

Tag Attributes

Tag attributes may be supplied as method arguments. Here is our previous example, but using attributes rather than tags:

@objects = [Object.new, Object.new, Object.new]

builder = Nokogiri::XML::Builder.new do |xml|
  xml.root {
    xml.objects {
      @objects.each do |o|
        xml.object(:type => o.type, :class => o.class, :id => o.id)
      end
    }
  }
end
puts builder.to_xml

Tag Attribute Short Cuts

A couple attribute short cuts are available when building tags. The short cuts are available by special method calls when building a tag.

This example builds an “object” tag with the class attribute “classy” and the id of “thing”:

builder = Nokogiri::XML::Builder.new do |xml|
  xml.root {
    xml.objects {
      xml.object.classy.thing!
    }
  }
end
puts builder.to_xml

Which will output:

<?xml version="1.0"?>
<root>
  <objects>
    <object class="classy" id="thing"/>
  </objects>
</root>

All other options are still supported with this syntax, including blocks and extra tag attributes.

Namespaces

Namespaces are added similarly to attributes. Nokogiri::XML::Builder assumes that when an attribute starts with “xmlns”, it is meant to be a namespace:

builder = Nokogiri::XML::Builder.new { |xml|
  xml.root('xmlns' => 'default', 'xmlns:foo' => 'bar') do
    xml.tenderlove
  end
}
puts builder.to_xml

Will output XML like this:

<?xml version="1.0"?>
<root xmlns:foo="bar" xmlns="default">
  <tenderlove/>
</root>

Referencing declared namespaces

Tags that reference non-default namespaces (i.e. a tag “foo:bar”) can be built by using the Nokogiri::XML::Builder#[] method.

For example:

builder = Nokogiri::XML::Builder.new do |xml|
  xml.root('xmlns:foo' => 'bar') {
    xml.objects {
      xml['foo'].object.classy.thing!
    }
  }
end
puts builder.to_xml

Will output this XML:

<?xml version="1.0"?>
<root xmlns:foo="bar">
  <objects>
    <foo:object class="classy" id="thing"/>
  </objects>
</root>

Note the “foo:object” tag.

Document Types

To create a document type (DTD), access use the Builder#doc method to get the current context document. Then call Node#create_internal_subset to create the DTD node.

For example, this Ruby:

builder = Nokogiri::XML::Builder.new do |xml|
  xml.doc.create_internal_subset(
    'html',
    "-//W3C//DTD HTML 4.01 Transitional//EN",
    "http://www.w3.org/TR/html4/loose.dtd"
  )
  xml.root do
    xml.foo
  end
end

puts builder.to_xml

Will output this xml:

<?xml version="1.0"?>
<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
<root>
  <foo/>
</root>