Research if usage of Lucene ranks is an option for 2 names similarity match

During on of the previous subtickets source code was refactored to use Lucene with a getSimilarity search that has two names as arguments.

PR: https://github.com/openmrs/openmrs-core/pull/3152

During the three name refactoring it was learned that the Lucene internal term rank gives a good enough result set by at the same time providing less and cleaner code. The discussion regarding this can be found here: https://talk.openmrs.org/t/soundex-search-in-lucenequeries/26068/39?u=fruether

In this ticket it should be investigated if the ranking method could as well be used for the two name search. If the result is that this would be feasible it can replace the previous code.