Concepts Extractor

Goal

This pipeline extensibility stage extract document concepts (from Fast docvector) and add them into a "concepts" crawled property. This allow users to refine on concetps, for example using our Tag Cloud Refiner webpart.

Download & Install

Download binairies here: http://spplayground.codeplex.com/releases/view/77688.
Put ConceptsExtractor.exe in c:\FastSearch\bin (depend of your installation), then modify the c:\FastSearch\etc\pipelineextensibility.xml file and add the content of the config.xml file:
<Run command="ConceptExtractor.exe %(input)s %(output)s">
  <Input>
    <CrawledProperty propertySet="48385C54-CDFC-4E84-8117-C95B3CF8911C" varType="31" propertyName="docvector"/>
  </Input>
  <Output>
    <CrawledProperty propertySet="fa585f53-2679-48d9-976d-9ce62e7e19b7" varType="31" propertyName="concepts"/>
  </Output>
</Run>


Once a full crawl has been done, create the managed property manually or use the install.ps1 file to do so.
Recrawl.

Voila.

Last edited Feb 27, 2012 at 3:12 PM by anotherspguy, version 4

Comments

No comments yet.