Content Ranking

[unknown-user]

Hi,

Came across this cool dashlet which is part of the Dashlet Challenge 2012. 
http://www.nabucco-source.org/confluence/display/MODALF/Content+Trends

We are also looking at implementing a similar type of content ranking but wanted it to be incorporated with Search. Basically, the search results needs to be ordered based on content rank which is decided by the following attributes.

Immediacy: How new is the item, how recently updated, how recently used
Frequency: Frequently accessed
Popularity: Most used, Most downloaded, Most recommended

I understand that there are ways to track the above attributes but how can we incorporate it with Solr search engine so that the results are shown based on the above attributes?

Thanks

[afaust]

Hello,

I'm glad you like the dashlet.

In order for SOLR to support sorting on your criteria, you need to find a way to derive sortable property values from the different criteria and maintain them at the node properties level. The main issue with time-related criteria is that you kind of have to recalculate the scoring of all your content in pre-defined intervals so that you have a consistent sort order.

If you have a closer look at the Content Trends dashlet, you'll see that I track each calculated score (total score, comment score, edit score …) as a property on the scored node, and use these properties to provide the sort via the search service (which can be either SOLR or Lucene). With some additional configuration, it could also be used in the regular search / document library.

Regards
Axel