We're using Lucene/SOLR and we want to display a distribution graph of relevancy scores for all results. We're thinking iterating through all the results would be very inefficient, any suggestions how to do this 'on the fly'? The main purpose is to allow users to remove results that are deemed of low relevancy (or we might do it for them) - any other solutions out there?
Asked
Active
Viewed 50 times
1
-
the main idea of collector in Lucene is that it only scores enough documents to maintain top-K results that you've asked initally (lets say 10, page 1), going through all results to score them all would be equally difficult even to collect the distribution of scores – Mysterion Oct 11 '19 at 11:09