CODEWITHSUNDEEP
lucene
Bruce Lowe
Bruce Lowe

Reputation: 6213

Lucene phrase match with a wildcard at the end

I'm trying to make a predictive text search that allows a user to start typing, and results from their text come through as results

E.g. with "ca" they can get "cat in the hat", "my calculus is cool" "cat dog mouse"

However, if a person continues to type with spaces, I want the whole phrase to be considered as one term

E.g. "cat i" should find "cat in the hat"

but NOT "[cat] dog mouse" nor "my calculus [i]s cool"

This is my current code, however it does not seem to be working as I'd hoped:

val mySort = new Sort(SortField.FIELD_SCORE, new SortField("popularity", SortField.Type.INT, true))
val analyzer = new StandardAnalyzer(Version.LUCENE_43)

val parser: QueryParser = new QueryParser(Version.LUCENE_43, "title", analyzer)
val query = parser.parse(queryString+"*")
val titleQuery = new ConstantScoreQuery(query)
titleQuery.setBoost(2)

val synopsisQuery = new QueryParser(Version.LUCENE_43, "synopsis", analyzer).parse(queryString)
val summaryQuery = new ConstantScoreQuery(synopsisQuery)

val finalQuery = new DisjunctionMaxQuery(0)
finalQuery.add(titleQuery)
finalQuery.add(summaryQuery)

val collector = TopFieldCollector.create(mySort,Limit,false,true,true,false)

searcher.search(finalQuery, collector)

collector.topDocs().scoreDocs

Upvotes: 1

Views: 757

Answers (1)

Lucas Trzesniewski
Lucas Trzesniewski

Reputation: 51430

There are basically two ways to achieve this.

The old way, is to construct a MultiPhraseQuery manually - see this answer for details.

The new way is simpler though: construct a SpanNearQuery. Use the following parameters: inOrder = true and slop = 0 to get an equivalent of PhraseQuery.

Each clause in the SpanNearQuery should be a SpanTermQuery except the last one. These should be the full terms contained in your phrase.

The last clause should be a SpanMultiTermQueryWrapper<PrefixQuery>, wrapping a PrefixQuery. Use the last term of your phrase as the prefix value.

To sum up, for cat i:

SpanNearQuery [inOrder = true, slop = 0]
 |
 +-- SpanTermQuery [term = "cat"]
 |
 +-- SpanMultiTermQueryWrapper<PrefixQuery>
      |
      +-- Prefixquery [prefix = "i"]

Upvotes: 4

Related Questions