Examples of Rescorer

org.apache.lucene.search.Rescorer
Re-scores the topN results ( {@link TopDocs}) from an original query. See {@link QueryRescorer} for an actualimplementation. Typically, you run a low-cost first-pass query across the entire index, collecting the top few hundred hits perhaps, and then use this class to mix in a more costly second pass scoring.
See {@link QueryRescorer#rescore(IndexSearcher,TopDocs,Query,double,int)}for a simple static method to call to rescore using a 2nd pass {@link Query}. @lucene.experimental
org.elasticsearch.search.rescore.Rescorer
A query rescorer interface used to re-rank the Top-K results of a previously executed search.

Examples of org.apache.lucene.search.Rescorer


    Expression e = JavascriptCompiler.compile("sqrt(_score) + ln(popularity)");
    SimpleBindings bindings = new SimpleBindings();
    bindings.add(new SortField("popularity", SortField.Type.INT));
    bindings.add(new SortField("_score", SortField.Type.SCORE));
    Rescorer rescorer = e.getRescorer(bindings);


    hits = rescorer.rescore(searcher, hits, 10);
    assertEquals(3, hits.totalHits);
    assertEquals("2", r.document(hits.scoreDocs[0].doc).get("id"));
    assertEquals("1", r.document(hits.scoreDocs[1].doc).get("id"));
    assertEquals("3", r.document(hits.scoreDocs[2].doc).get("id"));


    String expl = rescorer.explain(searcher,
                                   searcher.explain(query, hits.scoreDocs[0].doc),
                                   hits.scoreDocs[0].doc).toString();


    // Confirm the explanation breaks out the individual
    // variables:

View Full Code Here

Examples of org.elasticsearch.search.rescore.Rescorer

            context.parsedQuery(indexService.queryParserService().parseQuery(request.source()));
            context.preProcess();
            int topLevelDocId = result.docIdAndVersion().docId + result.docIdAndVersion().context.docBase;
            Explanation explanation = context.searcher().explain(context.query(), topLevelDocId);
            for (RescoreSearchContext ctx : context.rescore()) {
                Rescorer rescorer = ctx.rescorer();
                explanation = rescorer.explain(topLevelDocId, context, ctx, explanation);
            }
            if (request.fields() != null || (request.fetchSourceContext() != null && request.fetchSourceContext().fetchSource())) {
                // Advantage is that we're not opening a second searcher to retrieve the _source. Also
                // because we are working in the same searcher in engineGetResult we can be sure that a
                // doc isn't deleted between the initial get and this call.

View Full Code Here

TOP

All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.