List of usage examples for org.apache.lucene.search IndexSearcher createWeight
public Weight createWeight(Query query, ScoreMode scoreMode, float boost) throws IOException
From source file:com.o19s.es.explore.ExplorerQuery.java
License:Apache License
@Override public Weight createWeight(IndexSearcher searcher, boolean needsScores, float boost) throws IOException { if (!needsScores) { return searcher.createWeight(query, false, boost); }//from w w w .j a v a 2s. c om final Weight subWeight = searcher.createWeight(query, true, boost); Set<Term> terms = new HashSet<>(); subWeight.extractTerms(terms); if (isCollectionScoped()) { ClassicSimilarity sim = new ClassicSimilarity(); StatisticsHelper df_stats = new StatisticsHelper(); StatisticsHelper idf_stats = new StatisticsHelper(); StatisticsHelper ttf_stats = new StatisticsHelper(); for (Term term : terms) { TermContext ctx = TermContext.build(searcher.getTopReaderContext(), term); TermStatistics tStats = searcher.termStatistics(term, ctx); df_stats.add(tStats.docFreq()); idf_stats.add(sim.idf(tStats.docFreq(), searcher.getIndexReader().numDocs())); ttf_stats.add(tStats.totalTermFreq()); } /* If no terms are parsed in the query we opt for returning 0 instead of throwing an exception that could break various pipelines. */ float constantScore; if (terms.size() > 0) { switch (type) { case ("sum_classic_idf"): constantScore = idf_stats.getSum(); break; case ("mean_classic_idf"): constantScore = idf_stats.getMean(); break; case ("max_classic_idf"): constantScore = idf_stats.getMax(); break; case ("min_classic_idf"): constantScore = idf_stats.getMin(); break; case ("stddev_classic_idf"): constantScore = idf_stats.getStdDev(); break; case "sum_raw_df": constantScore = df_stats.getSum(); break; case "min_raw_df": constantScore = df_stats.getMin(); break; case "max_raw_df": constantScore = df_stats.getMax(); break; case "mean_raw_df": constantScore = df_stats.getMean(); break; case "stddev_raw_df": constantScore = df_stats.getStdDev(); break; case "sum_raw_ttf": constantScore = ttf_stats.getSum(); break; case "min_raw_ttf": constantScore = ttf_stats.getMin(); break; case "max_raw_ttf": constantScore = ttf_stats.getMax(); break; case "mean_raw_ttf": constantScore = ttf_stats.getMean(); break; case "stddev_raw_ttf": constantScore = ttf_stats.getStdDev(); break; case "unique_terms_count": constantScore = terms.size(); break; default: throw new RuntimeException("Invalid stat type specified."); } } else { constantScore = 0.0f; } return new ConstantScoreWeight(ExplorerQuery.this, constantScore) { @Override public Explanation explain(LeafReaderContext context, int doc) throws IOException { Scorer scorer = scorer(context); int newDoc = scorer.iterator().advance(doc); assert newDoc == doc; // this is a DocIdSetIterator.all return Explanation.match(scorer.score(), "Stat Score: " + type); } @Override public Scorer scorer(LeafReaderContext context) throws IOException { return new ConstantScoreScorer(this, constantScore, DocIdSetIterator.all(context.reader().maxDoc())); } @Override public boolean isCacheable(LeafReaderContext ctx) { return true; } }; } else if (type.endsWith("_raw_tf")) { // Rewrite this into a boolean query where we can inject our PostingsExplorerQuery BooleanQuery.Builder qb = new BooleanQuery.Builder(); for (Term t : terms) { qb.add(new BooleanClause(new PostingsExplorerQuery(t, PostingsExplorerQuery.Type.TF), BooleanClause.Occur.SHOULD)); } // FIXME: completely refactor this class and stop accepting a random query but a list of terms directly // rewriting at this point is wrong, additionally we certainly build the TermContext twice for every terms // problem is that we rely on extractTerms which happen too late in the process Query q = qb.build().rewrite(searcher.getIndexReader()); return new ExplorerQuery.ExplorerWeight(this, searcher.createWeight(q, true, boost), type); } throw new IllegalArgumentException("Unknown ExplorerQuery type [" + type + "]"); }
From source file:com.o19s.es.ltr.query.RankerQuery.java
License:Apache License
@Override public Weight createWeight(IndexSearcher searcher, boolean needsScores, float boost) throws IOException { if (!needsScores) { // If scores are not needed simply return a constant score on all docs return new ConstantScoreWeight(this, boost) { @Override/*from w w w. jav a2 s . com*/ public Scorer scorer(LeafReaderContext context) throws IOException { return new ConstantScoreScorer(this, score(), DocIdSetIterator.all(context.reader().maxDoc())); } @Override public boolean isCacheable(LeafReaderContext ctx) { return true; } }; } List<Weight> weights = new ArrayList<>(queries.size()); for (Query q : queries) { weights.add(searcher.createWeight(q, needsScores, boost)); } return new RankerWeight(weights); }
From source file:org.tallison.lucene.search.concordance.charoffsets.SpansCrawler.java
License:Apache License
public static void crawl(SpanQuery query, Query filter, IndexSearcher searcher, DocTokenOffsetsVisitor visitor) throws IOException, TargetTokenNotFoundException { query = (SpanQuery) query.rewrite(searcher.getIndexReader()); SpanWeight w = query.createWeight(searcher, ScoreMode.COMPLETE_NO_SCORES, 1.0f); if (filter == null) { for (LeafReaderContext ctx : searcher.getIndexReader().leaves()) { Spans spans = w.getSpans(ctx, SpanWeight.Postings.POSITIONS); if (spans == null) { continue; }//from www.j av a2 s . c o m boolean cont = visitLeafReader(ctx, spans, visitor); if (!cont) { break; } } } else { filter = searcher.rewrite(filter); Weight searcherWeight = searcher.createWeight(filter, ScoreMode.COMPLETE_NO_SCORES, 1.0f); for (LeafReaderContext ctx : searcher.getIndexReader().leaves()) { Scorer leafReaderContextScorer = searcherWeight.scorer(ctx); if (leafReaderContextScorer == null) { continue; } //Can we tell from the scorer that there were no hits? //in <= 5.x we could stop here if the filter query had no hits. Spans spans = w.getSpans(ctx, SpanWeight.Postings.POSITIONS); if (spans == null) { continue; } DocIdSetIterator filterItr = leafReaderContextScorer.iterator(); if (filterItr == null || filterItr.equals(DocIdSetIterator.empty())) { continue; } boolean cont = visitLeafReader(ctx, spans, filterItr, visitor); if (!cont) { break; } } } }
From source file:tech.beshu.ror.es.security.DocumentFilterReader.java
License:Open Source License
private DocumentFilterReader(LeafReader reader, Query query) throws IOException { super(reader); final IndexSearcher searcher = new IndexSearcher(this); searcher.setQueryCache(null);//w w w. j a v a 2 s . c o m final boolean needsScores = false; final Weight preserveWeight = searcher.createWeight(query, needsScores, 0); final int maxDoc = this.in.maxDoc(); final FixedBitSet bits = new FixedBitSet(maxDoc); final Scorer preserveScorer = preserveWeight.scorer(this.getContext()); if (preserveScorer != null) { bits.or(preserveScorer.iterator()); } if (in.hasDeletions()) { final Bits oldLiveDocs = in.getLiveDocs(); assert oldLiveDocs != null; final DocIdSetIterator it = new BitSetIterator(bits, 0L); for (int i = it.nextDoc(); i != DocIdSetIterator.NO_MORE_DOCS; i = it.nextDoc()) { if (!oldLiveDocs.get(i)) { bits.clear(i); } } } this.liveDocs = bits; this.numDocs = bits.cardinality(); }