Can you please elaborate. As an example, I need to evaluate the query "albert einstein" with result documents having the following score:
Score("albert einstein") = w1 * score(albert) + w2*score(einstein) + w3 ((albert einstein)~1) -- the last is a proximity query.
I guess I would need to:
1- create a custom parser (a parser emits a Query)
2- add a Scorer on top of each term (how would this combine with the output of the parser, which I guess would emit a MultiTerm query?)
3- use the LinearAggregator to combine scores from the different Scorers with the appropriate.
I guess the confusion comes from 1 and 2.