SearchMonkey Guide

The SearchMonkey vocabularies is a collection of vocabularies that we recommend using in DataRSS feeds and for annotating pages with metadata. Each vocabulary includes a set of terms and classes that are common for a particular domain. The set of vocabularies includes well-established RDF vocabularies as well as RDF vocabularies for microformats such as hCard, hCalendar and hReview. The vocabulary set also includes vocabularies developed specifically for SearchMonkey.

The following table lists the standard vocabularies recommended to be used in SearchMonkey applications. The documentation for some of these vocabularies is reproduced in this guide for convenience.

These vocabularies are intended to help developers to get started. However, this selection is not exclusive: you can provide data using other vocabularies. See Defining New Properties for more information.

The section Predefined Prefixes gives an overview of the recommended vocabularies. In dataRSS feeds, these vocabularies do not need to be explicitly declared if you place the following processing instruction at the beginning of a dataRSS feed:

<?profile http://search.yahoo.com/searchmonkey-profile ?>

In pages with embedded RDF metadata (eRDF and RDFa), each vocabulary needs to be declared using the appropriate constructs, i.e. LINK elements in eRDF and XML namespace declarations in RDFa.

In section Examples we list a number of examples of using these vocabularies in DataRSS.

This specification is on a periodic release schedule to improve conformance to industry standard vocabularies and enable common use cases to be accomplished in a consistent manner.

In the following we include some examples of representing data in DataRSS format from different domains. (We ignore the Atom headers for brevity). As DataRSS follows the RDFa standard, these examples can be directly translated to annotations in HTML by applying the same attributes (rel, property, typeof, resource) to HTML elements, following the same nesting as in the examples.

As an example, consider the following snippet of dataRSS:

          <y:item rel="dc:subject">
            <y:type typeof="foaf:Person>
              <y:meta property="foaf:name">John Doe</y:meta>
              <y:item rel="foaf:homepage" resource="http://www.joeisageek.com"/>
              <y:item rel="foaf:knows">
                <type typeof="foaf:Person">
                  <y:meta property="foaf:name">Jane Doe</y:meta>
                  <y:meta property="foaf:gender">female</y:meta>
                  <y:item rel="foaf:mbox" resource="mailto:janedoe@example.org"/>
                </type>
              </y:item>
            </y:type>
          </y:item>
        

This can be directly embedded inside HTML by applying the attribues to non-display HTML elements such as SPAN and DIV:

          <span rel="dc:subject">
            <span typeof="foaf:Person>
              <span property="foaf:name">John Doe</span>
              <span rel="foaf:homepage" resource="http://www.joeisageek.com"/>
              <span rel="foaf:knows">
                <span typeof="foaf:Person">
                  <span property="foaf:name">Jane Doe</span>
                  <span property="foaf:gender">female</span>
                  <span rel="foaf:mbox" resource="mailto:janedoe@example.org"/>
                </span>
              </span>
            </span>
          </span>
        

Note that RDFa provides additional attributes that make it easier to add markup to existing pages by reusing semantic-bearing HTML elements. For example, in the above case both the name and homepage can be provided using the HTML <a> tag:

          <a  rel="foaf:homepage" href="http://www.joeisageek.com" property="foaf:name" >John Doe</a>
        

See the RDFa Primer for more details.

There will probably be times when the kinds of metadata you'd like to extract isn't found in the searchmonkey-profile vocabulary reference. It might be an existing RDF vocabulary, or something you needed to make up by yourself.

For example, suppose your metadata is about digital cameras and in particular you would like to represent the number of megapixels a digital camera can handle. Let's assume you have created an RDF or OWL ontology and defined the term DigitalCamera and the property megapixels, both with a namespace http://example.com/vocab/digicam#. You are encouraged to publish your schema at this same location so that others can consult the definition of your newly created term.

The declaration and use of your terms in combination with the existing class product:Product would look like this:

<y:adjunct id="1.0" name="plisting" xmlns:digicam="http://example.com/vocab/digicam#">
  <y:item rel="product:Product">
     <y:item rel="rdf:type" resource="http://example.com/vocab/digicam#DigitalCamera"/>
     <y:meta prop="digicam:megapixels">7.2</y:meta>
  </y:item>
</y:adjunct>

Use the links below to download the OWL definitions of the SearchMonkey vocabularies.

Click here for more Yahoo! published OWL vocabulary definitions.

The following table lists the prefixes that are predefined for DataRSS feeds. Namespacesthat are not included in this table need to be explicitly defined. It is also an error to redefine any of these namespaces.

Datatype vocabularies are used with in combination with literals to specify the type of literal. Some properties, like dc:identifier, may have many different possible values. To specify the type of the actual value provide, use the datatype attribute. In the following example, a datatype attribute specifies that the value of the dc:identifier is a positive integer.

	Note
	The datatype attribute is supported using the data::xpath , you cannot query it using data::get.

<y:meta property="dc:identifier" datatype="xsd:positiveInteger">
0545010225 </y:meta>

Datatypes help to validate the output of custom data services: a warning can be raisedif the actual value does not conform to the specified datatype.