Setting Metadata Automatically

HARP automatically gathers metadata on XML content when it is added to a project using XSLT stylesheets that output a simple XML document describing the metadata on the object.

To create custom metadata rules, create a file called harp-metadata.xsl in a folder called HARP-META at the root level of a project or document type. Metadata rules for a doctype will apply for all documents of that doctype in all projects using that doctype. Rules in a project apply for all XML documents in that project. If a file has metadata rules associated with both its project and its doctype, both transforms will be run, and the union of all metadata gathered from both will be assigned to the object.

The stylesheet should output XML in the following format:

<root>
  <metadata name="name">value</metadata>
  <!-- Other metadata -->
</root>

If a given metadata name has more than one value, the XSLT should output multiple <metadata> elements with that name. Any markup within <metadata> elements will be ignored; only character data will be used.

Default Metadata Rules for DITA Documents

The built-in DITA doctypes are configured with the following metadata rules. The default stylesheet will work for the given element names as well as any specializations of those elements.

product: The @product attribute on the root element.; The <prodname> elements within <prodinfo> elements within <prolog> (on topics) or top-level <topicmeta> (on maps).
audience: The @audience attribute on the root element.; The @type attribute on <audience> elements within <prolog> (on topics) or top-level <topicmeta> (on maps).
experiencelevel: The @expriencelevel attribute on <audience> elements within <prolog> (on topics) or top-level <topicmeta> (on maps).
platform: The @platform attribute on the root element.; The <platform> elements within <prodinfo> elements with <prolog> (on topics) or top-level <topicmeta> (on maps).
brand: The <brand> elements within <prodinfo> elements within <prolog> (on topics) or top-level <topicmeta> (on maps).
permission: The @view attribute on any <permission> elements within <prolog> (on topics) or top-level <topicmeta> (on maps).
category: Any <category> elements within <prolog> (on topics) or top-level <topicmeta> (on maps).

The default rules will also capture <data> elements wtithin <metadata> elements within <prolog> (on topics) or top-level <topicmeta> (on maps). The metadata name is drawn from the @name attribute, and the value is drawn from the @value attribute and the element contents. If both the @value attribute is present and the element contains content, two values will be assigned to the metadata name.

Finally, the default rules will automatically capture <othermeta> elements within <prolog> (on topics) or top-level <topicmeta> (on maps). The name will be drawn from the @name attribute, and the value from @content.