Related Results with Carrot
From KallestadWiki
What is Carrot2
Carrot2 is a Search Results Clustering Engine. What it does is automatically organize result sets within thematic categories, based on known information about the results. The best way to understand what Carrot can do is to see the results live. Head over to the Carrot2 Demo application and give it a whirl.
Combining Carrot with Lucene
Carrot has an API specifically for Lucene integrations. It requires three fields within Lucene, or programatically discoverable:
- Unique Identifier (url)
- A Document Title
- A Snippet of the actual document
Carrot filters through the result set of a given query, clusters the results together, and outputs a final data set. A code example is available at sourceforge.
How you implement this functionality is up to you. You could have carrot stand in front of Lucene, have it parse your query, and output a web page as a result - or you could have the whole thing tied together behind the scenes while your own custom application handles query input and results output. The power that carrot provides is in it's ability to cluster, and the ability to assign different clustering algorithms. One possibility is to track your popular search criteria, and then to analyze the carrot results behind the scenes. You could insert static results that are more refined than carrot can manage dynamically, and you would gain a performance benefit in outputting static data.
Full Text Search Index
- About Full Text Searching
- Lucene
- Lucene Documents
- Lucene Indexes
- Integration Points with Lucene
- Related Results with Carrot
