public override Query Rewrite(IndexReader reader)
{
if (rewrittenQuery != null)
{
return(rewrittenQuery);
}
//load up the list of possible terms
foreach (FieldVals f in fieldVals)
{
AddTerms(reader, f);
}
//clear the list of fields
fieldVals.Clear();
BooleanQuery bq = new BooleanQuery();
//create BooleanQueries to hold the variants for each token/field pair and ensure it
// has no coord factor
//Step 1: sort the termqueries by term/field
HashMap <Term, List <ScoreTerm> > variantQueries = new HashMap <Term, List <ScoreTerm> >();
int size = q.Size();
for (int i = 0; i < size; i++)
{
ScoreTerm st = q.Pop();
var l = variantQueries[st.fuzziedSourceTerm];
if (l == null)
{
l = new List <ScoreTerm>();
variantQueries.Add(st.fuzziedSourceTerm, l);
}
l.Add(st);
}
//Step 2: Organize the sorted termqueries into zero-coord scoring boolean queries
foreach (var variants in variantQueries.Values)
{
if (variants.Count == 1)
{
//optimize where only one selected variant
ScoreTerm st = variants[0];
TermQuery tq = new FuzzyTermQuery(st.Term, ignoreTF);
tq.Boost = st.Score; // set the boost to a mix of IDF and score
bq.Add(tq, Occur.SHOULD);
}
else
{
BooleanQuery termVariants = new BooleanQuery(true); //disable coord and IDF for these term variants
foreach (ScoreTerm st in variants)
{
TermQuery tq = new FuzzyTermQuery(st.Term, ignoreTF); // found a match
tq.Boost = st.Score; // set the boost using the ScoreTerm's score
termVariants.Add(tq, Occur.SHOULD); // add to query
}
bq.Add(termVariants, Occur.SHOULD); // add to query
}
}
//TODO possible alternative step 3 - organize above booleans into a new layer of field-based
// booleans with a minimum-should-match of NumFields-1?
bq.Boost = Boost;
this.rewrittenQuery = bq;
return(bq);
}
