Although issue #38 guarantees correct counts for the heuristically determined terms, it is possible to turn off this feature. Doing so should scale up the counts, based on the fraction of the result set size used for sampling. This would in effect implement issue #37.
Although issue #38 guarantees correct counts for the heuristically determined terms, it is possible to turn off this feature. Doing so should scale up the counts, based on the fraction of the result set size used for sampling. This would in effect implement issue #37.