RDF

A general guideline on how to interpret the Calais response in RDF is presented here.

For extracted metadata elements (Entities, Events and Facts) the RDF includes the following artifacts:

Examples:

The entity Company for "ClearForest Ltd." may look like this:

<rdf:Description rdf:about="http://d.opencalais.com/comphash-1/9dd2192a-4cd2-3b9a-ac2f-b6a0d1fed773">
<rdf:type rdf:resource="http://s.opencalais.com/1/type/em/e/Company"/>
<c:name>ClearForest Ltd.</c:name>
</rdf:Description>

The event Acquisition between "Reuters" and “ClearForest Ltd.” may look like this:

<rdf:Description rdf:about="http://d.opencalais.com/genericHasher-1/e83cd693-2146-32a2-b1fe-c4a73615dbf0">
 <rdf:type rdf:resource="http://s.opencalais.com/1/type/em/r/Acquisition"/>
 <!--Reuters-->
 <c:company_acquirer rdf:resource="http://d.opencalais.com/comphash-1/48344864-ce62-3064-ae05-a3b41fab186c"/>
 <!--ClearForest Ltd.-->
 <c:company_beingacquired rdf:resource="http://d.opencalais.com/comphash-1/9dd2192a-4cd2-3b9a-ac2f-b6a0d1fed773"/>
 <c:status>planned</c:status>
</rdf:Description>

Examples:

An instance for the unique company ClearForest Ltd. may look like this:

<rdf:Description rdf:about="http://d.opencalais.com/dochash-1/37042f2e-8a72-30c7-afe2-2710efcb4d5f/Instance/11">
 <rdf:type rdf:resource="http://s.opencalais.com/1/type/sys/InstanceInfo"/>
 <c:docId rdf:resource="http://d.opencalais.com/dochash-1/37042f2e-8a72-30c7-afe2-2710efcb4d5f"/>
 <c:subject rdf:resource="http://d.opencalais.com/comphash-1/9dd2192a-4cd2-3b9a-ac2f-b6a0d1fed773"/>
 <!--Company: ClearForest Ltd.-->
 <c:detection>[Reuters to acquire text search firm ]ClearForest[ &lt;/TITLE&gt;
 &lt;DATE&gt; Mon Apr 30, 2007 7:00am EDT]</c:detection>
 <c:offset>55</c:offset>
 <c:length>11</c:length>
 </rdf:Description>

An instance for the unique Acquisition element between "Reuters" and "ClearForest Ltd." may look like this:

<rdf:Description rdf:about="http://d.opencalais.com/dochash-1/37042f2e-8a72-30c7-afe2-2710efcb4d5f/Instance/22">
 <rdf:type rdf:resource="http://s.opencalais.com/1/type/sys/InstanceInfo"/>
 <c:docId rdf:resource="http://d.opencalais.com/dochash-1/37042f2e-8a72-30c7-afe2-2710efcb4d5f"/>
 <c:subject rdf:resource="http://d.opencalais.com/genericHasher-1/e83cd693-2146-32a2-b1fe-c4a73615dbf0"/>
 <!--Acquisition: company_acquirer: Reuters; company_beingacquired: ClearForest Ltd.; status: planned; -->
 <c:detection>[&lt;DOCUMENT&gt;
 &lt;TITLE&gt; ]Reuters to acquire text search firm ClearForest[ &lt;/TITLE&gt;
 &lt;DATE&gt; Mon Apr 30, 2007 7:00am EDT]</c:detection>
 <c:offset>19</c:offset>
 <c:length>47</c:length>
 </rdf:Description>

In addition, the RDF response includes general document and transaction information, such as the document language, submission date and time and request ID. It also includes the input content after it has been converted into valid XML for the actual processing by the Calais backend server (except for the TEXT/RAW option).

There are several changes and enhancements related to the RDF output format. Users asked for an option to exclude the original body from the returned RDF output. OpenCalais now supports this option through a new parameter in the processing directives section of paramsXML:

c:omitOutputtingOriginalText="TRUE"

By default, the original body is returned in any RDF response.

Another change in RDF output is that the RDF header, which includes a summary of all entities extracted from the text, is now sorted alphabetically based on the entity type (the same sorting used in Simple Format).

Lastly, disambiguation results are integrated in RDF output in the following manner.

For Companies: Resolution nodes are added to the output RDF. Each such node contains the following information:

• c:subject - URI of the referred company entity. A resolution node may contain multiple subject properties; one for each company entity which was resolved to this single company.
• c:docId – URI of the document this resolution was created in.
• c:score – a score representing the certainty with which the company was resolved.
• c:name – formal English legal name of resolved company.
• c:ticker – company’s ticker.

The RDF example below shows a company entity (top RDF node) and the respective resolution node.

<rdf:Description rdf:about="http://d.opencalais.com/comphash-1/64136b2b-cb4e-36ac-9f32-f58f4c1f1c8a">
  <rdf:type rdf:resource="http://s.opencalais.com/1/type/em/e/Company" />
  <c:name>British Airways</c:name>
  <c:nationality>British</c:nationality>
</rdf:Description>
<rdf:Description rdf:about="http://d.opencalais.com/er/company/ralg-tr1r/58ad4ecb-2df0-3d46-8333-2d25dcb364d9">
  <rdf:type rdf:resource="http://s.opencalais.com/1/type/er/Company" />
  <c:docId rdf:resource="http://d.opencalais.com/dochash-1/88096fc6-9ea2-3c9f-a0a0-c29a0a5fdced" /> 
  <c:subject rdf:resource="http://d.opencalais.com/comphash-1/64136b2b-cb4e-36ac-9f32-f58f4c1f1c8a" />
  <c:score>1.0</c:score>
  <c:name>British Airways PLC</c:name>
  <c:ticker>BAY</c:ticker>
</rdf:Description>

For Geographies: Resolution nodes will be added to the output RDF. Each such node contains the ID of the referred entity (city or province or state or country found in input) in its "c:subject" property, the resolved name in its "c:name" property, latitude of the location in its "c:lat" property and longitude of the location in its "c:long" property.

The RDF example below shows a city entity (top RDF node) and the respective resolution node.

<rdf:Description rdf:about="http://d.opencalais.com/genericHasher-1/96e9e28b-f95c-3f9c-a374-b3bcfbc02cfd">
  <rdf:type rdf:resource="http://s.opencalais.com/1/type/em/e/City"/>
  <c:name>Golden</c:name>
</rdf:Description>
<rdf:Description rdf:about="http://d.opencalais.com/er/geo/ralg-geo1/e3f5b88c-f2f2-6e4f-7e2c-f0452221c341">
  <rdf:type rdf:resource="http://s.opencalais.com/1/type/er/Geo"/>
  <c:docId rdf:resource="http://d.opencalais.com/dochash-1/3508bef0-f669-3dec-829c-d3344507f857"/>
  <!--Golden-->
  <c:subject rdf:resource="http://d.opencalais.com/genericHasher-1/96e9e28b-f95c-3f9c-a374-b3bcfbc02cfd"/>>
  <c:name>Golden,Colorado,United States</c:name>
  <c:lat>39.7556</c:lat>
  <c:long>-105.2206</c:long>
</rdf:Description>

A timeout of 20 seconds is applied if large input content is submitted to Calais., however, instead of dropping the transaction, Calais will return the metadata results extracted so far, and will also indicate the occurrence of a timeout for the submitted content. The message will be as follows:

<c:message>
  <rdf:Description>
  <rdf:type rdf:resource="http://s.opencalais.com/1/type/sys/Message" />
   <c:messageCode>201</c:messageCode>
   <c:text>Partial metadata extraction due to timeout </c:text>
 </rdf:Description>
</c:message>

Back to Top

Examples

Attached are two example files: input in the form of a TEXT/XML document submitted to Calais, and its resulting RDF output file.

Back to Top