Smarter approach to query processing

[searchivarius]

Currently we expect only white-space tokenized queries. However they can accidentally contain special symbols and operators in Lucene. These can be potentially indexed sometimes. We filter them out during querying, but a smarter approach would consist in formulating the query directly using the BooleanQuery and TermQuery: https://www.tutorialspoint.com/lucene/lucene_booleanquery.htm

If we every support Lucene analyzers, however, respective queries may also contain Lucene operators. How are we going to deal with it? Perhaps, we need to see how it's done in the Elastic Search.