All Downloads are FREE. Search and download functionalities are using the official Maven repository.

org.elasticsearch.common.lucene.search.function.FiltersFunctionScoreQuery Maven / Gradle / Ivy

There is a newer version: 8.14.1
Show newest version
/*
 * Licensed to Elasticsearch under one or more contributor
 * license agreements. See the NOTICE file distributed with
 * this work for additional information regarding copyright
 * ownership. Elasticsearch licenses this file to you under
 * the Apache License, Version 2.0 (the "License"); you may
 * not use this file except in compliance with the License.
 * You may obtain a copy of the License at
 *
 *    http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing,
 * software distributed under the License is distributed on an
 * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
 * KIND, either express or implied.  See the License for the
 * specific language governing permissions and limitations
 * under the License.
 */

package org.elasticsearch.common.lucene.search.function;

import org.apache.lucene.index.IndexReader;
import org.apache.lucene.index.LeafReaderContext;
import org.apache.lucene.index.Term;
import org.apache.lucene.search.Explanation;
import org.apache.lucene.search.FilterScorer;
import org.apache.lucene.search.IndexSearcher;
import org.apache.lucene.search.Query;
import org.apache.lucene.search.Scorer;
import org.apache.lucene.search.Weight;
import org.apache.lucene.util.Bits;
import org.elasticsearch.common.io.stream.StreamInput;
import org.elasticsearch.common.io.stream.StreamOutput;
import org.elasticsearch.common.io.stream.Writeable;
import org.elasticsearch.common.lucene.Lucene;

import java.io.IOException;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.Collection;
import java.util.Collections;
import java.util.List;
import java.util.Locale;
import java.util.Objects;
import java.util.Set;

/**
 * A query that allows for a pluggable boost function / filter. If it matches
 * the filter, it will be boosted by the formula.
 */
public class FiltersFunctionScoreQuery extends Query {

    public static class FilterFunction {
        public final Query filter;
        public final ScoreFunction function;

        public FilterFunction(Query filter, ScoreFunction function) {
            this.filter = filter;
            this.function = function;
        }

        @Override
        public boolean equals(Object o) {
            if (this == o) {
                return true;
            }
            if (o == null || getClass() != o.getClass()) {
                return false;
            }
            FilterFunction that = (FilterFunction) o;
            return Objects.equals(this.filter, that.filter) && Objects.equals(this.function, that.function);
        }

        @Override
        public int hashCode() {
            return Objects.hash(super.hashCode(), filter, function);
        }
    }

    public enum ScoreMode implements Writeable {
        FIRST, AVG, MAX, SUM, MIN, MULTIPLY;

        @Override
        public void writeTo(StreamOutput out) throws IOException {
            out.writeVInt(this.ordinal());
        }

        public static ScoreMode readFromStream(StreamInput in) throws IOException {
            int ordinal = in.readVInt();
            if (ordinal < 0 || ordinal >= values().length) {
                throw new IOException("Unknown ScoreMode ordinal [" + ordinal + "]");
            }
            return values()[ordinal];
        }

        public static ScoreMode fromString(String scoreMode) {
            return valueOf(scoreMode.toUpperCase(Locale.ROOT));
        }
    }

    final Query subQuery;
    final FilterFunction[] filterFunctions;
    final ScoreMode scoreMode;
    final float maxBoost;
    private final Float minScore;

    protected final CombineFunction combineFunction;

    public FiltersFunctionScoreQuery(Query subQuery, ScoreMode scoreMode, FilterFunction[] filterFunctions, float maxBoost, Float minScore, CombineFunction combineFunction) {
        this.subQuery = subQuery;
        this.scoreMode = scoreMode;
        this.filterFunctions = filterFunctions;
        this.maxBoost = maxBoost;
        this.combineFunction = combineFunction;
        this.minScore = minScore;
    }

    public Query getSubQuery() {
        return subQuery;
    }

    public FilterFunction[] getFilterFunctions() {
        return filterFunctions;
    }

    @Override
    public Query rewrite(IndexReader reader) throws IOException {
        Query rewritten = super.rewrite(reader);
        if (rewritten != this) {
            return rewritten;
        }
        Query newQ = subQuery.rewrite(reader);
        if (newQ == subQuery)
            return this;
        return new FiltersFunctionScoreQuery(newQ, scoreMode, filterFunctions, maxBoost, minScore, combineFunction);
    }

    @Override
    public Weight createWeight(IndexSearcher searcher, boolean needsScores) throws IOException {
        if (needsScores == false && minScore == null) {
            return subQuery.createWeight(searcher, needsScores);
        }

        boolean subQueryNeedsScores = combineFunction != CombineFunction.REPLACE;
        Weight[] filterWeights = new Weight[filterFunctions.length];
        for (int i = 0; i < filterFunctions.length; ++i) {
            subQueryNeedsScores |= filterFunctions[i].function.needsScores();
            filterWeights[i] = searcher.createNormalizedWeight(filterFunctions[i].filter, false);
        }
        Weight subQueryWeight = subQuery.createWeight(searcher, subQueryNeedsScores);
        return new CustomBoostFactorWeight(this, subQueryWeight, filterWeights, subQueryNeedsScores);
    }

    class CustomBoostFactorWeight extends Weight {

        final Weight subQueryWeight;
        final Weight[] filterWeights;
        final boolean needsScores;

        public CustomBoostFactorWeight(Query parent, Weight subQueryWeight, Weight[] filterWeights, boolean needsScores) throws IOException {
            super(parent);
            this.subQueryWeight = subQueryWeight;
            this.filterWeights = filterWeights;
            this.needsScores = needsScores;
        }

        @Override
        public void extractTerms(Set terms) {
            subQueryWeight.extractTerms(terms);
        }

        @Override
        public float getValueForNormalization() throws IOException {
            return subQueryWeight.getValueForNormalization();
        }

        @Override
        public void normalize(float norm, float boost) {
            subQueryWeight.normalize(norm, boost);
        }

        private FiltersFunctionFactorScorer functionScorer(LeafReaderContext context) throws IOException {
            Scorer subQueryScorer = subQueryWeight.scorer(context);
            if (subQueryScorer == null) {
                return null;
            }
            final LeafScoreFunction[] functions = new LeafScoreFunction[filterFunctions.length];
            final Bits[] docSets = new Bits[filterFunctions.length];
            for (int i = 0; i < filterFunctions.length; i++) {
                FilterFunction filterFunction = filterFunctions[i];
                functions[i] = filterFunction.function.getLeafScoreFunction(context);
                Scorer filterScorer = filterWeights[i].scorer(context);
                docSets[i] = Lucene.asSequentialAccessBits(context.reader().maxDoc(), filterScorer);
            }
            return new FiltersFunctionFactorScorer(this, subQueryScorer, scoreMode, filterFunctions, maxBoost, functions, docSets, combineFunction, needsScores);
        }

        @Override
        public Scorer scorer(LeafReaderContext context) throws IOException {
            Scorer scorer = functionScorer(context);
            if (scorer != null && minScore != null) {
                scorer = new MinScoreScorer(this, scorer, minScore);
            }
            return scorer;
        }

        @Override
        public Explanation explain(LeafReaderContext context, int doc) throws IOException {

            Explanation expl = subQueryWeight.explain(context, doc);
            if (!expl.isMatch()) {
                return expl;
            }
            // First: Gather explanations for all filters
            List filterExplanations = new ArrayList<>();
            for (int i = 0; i < filterFunctions.length; ++i) {
                Bits docSet = Lucene.asSequentialAccessBits(context.reader().maxDoc(),
                        filterWeights[i].scorer(context));
                if (docSet.get(doc)) {
                    FilterFunction filterFunction = filterFunctions[i];
                    Explanation functionExplanation = filterFunction.function.getLeafScoreFunction(context).explainScore(doc, expl);
                    double factor = functionExplanation.getValue();
                    float sc = CombineFunction.toFloat(factor);
                    Explanation filterExplanation = Explanation.match(sc, "function score, product of:",
                            Explanation.match(1.0f, "match filter: " + filterFunction.filter.toString()), functionExplanation);
                    filterExplanations.add(filterExplanation);
                }
            }
            FiltersFunctionFactorScorer scorer = functionScorer(context);
            int actualDoc = scorer.iterator().advance(doc);
            assert (actualDoc == doc);
            double score = scorer.computeScore(doc, expl.getValue());
            Explanation factorExplanation;
            if (filterExplanations.size() > 0) {
                factorExplanation = Explanation.match(
                        CombineFunction.toFloat(score),
                        "function score, score mode [" + scoreMode.toString().toLowerCase(Locale.ROOT) + "]",
                        filterExplanations);

            } else {
                // it is a little weird to add a match although no function matches but that is the way function_score behaves right now
                factorExplanation = Explanation.match(1.0f,
                    "No function matched", Collections.emptyList());
            }
            expl = combineFunction.explain(expl, factorExplanation, maxBoost);
            if (minScore != null && minScore > expl.getValue()) {
                expl = Explanation.noMatch("Score value is too low, expected at least " + minScore + " but got " + expl.getValue(), expl);
            }
            return expl;
        }
    }

    static class FiltersFunctionFactorScorer extends FilterScorer {
        private final FilterFunction[] filterFunctions;
        private final ScoreMode scoreMode;
        private final LeafScoreFunction[] functions;
        private final Bits[] docSets;
        private final CombineFunction scoreCombiner;
        private final float maxBoost;
        private final boolean needsScores;

        private FiltersFunctionFactorScorer(CustomBoostFactorWeight w, Scorer scorer, ScoreMode scoreMode, FilterFunction[] filterFunctions,
                                            float maxBoost, LeafScoreFunction[] functions, Bits[] docSets, CombineFunction scoreCombiner, boolean needsScores) throws IOException {
            super(scorer, w);
            this.scoreMode = scoreMode;
            this.filterFunctions = filterFunctions;
            this.functions = functions;
            this.docSets = docSets;
            this.scoreCombiner = scoreCombiner;
            this.maxBoost = maxBoost;
            this.needsScores = needsScores;
        }

        @Override
        public float score() throws IOException {
            int docId = docID();
            // Even if the weight is created with needsScores=false, it might
            // be costly to call score(), so we explicitly check if scores
            // are needed
            float subQueryScore = needsScores ? super.score() : 0f;
            double factor = computeScore(docId, subQueryScore);
            return scoreCombiner.combine(subQueryScore, factor, maxBoost);
        }

        protected double computeScore(int docId, float subQueryScore) {
            double factor = 1d;
            switch(scoreMode) {
                case FIRST:
                    for (int i = 0; i < filterFunctions.length; i++) {
                        if (docSets[i].get(docId)) {
                            factor = functions[i].score(docId, subQueryScore);
                            break;
                        }
                    }
                    break;
                case MAX:
                    double maxFactor = Double.NEGATIVE_INFINITY;
                    for (int i = 0; i < filterFunctions.length; i++) {
                        if (docSets[i].get(docId)) {
                            maxFactor = Math.max(functions[i].score(docId, subQueryScore), maxFactor);
                        }
                    }
                    if (maxFactor != Float.NEGATIVE_INFINITY) {
                        factor = maxFactor;
                    }
                    break;
                case MIN:
                    double minFactor = Double.POSITIVE_INFINITY;
                    for (int i = 0; i < filterFunctions.length; i++) {
                        if (docSets[i].get(docId)) {
                            minFactor = Math.min(functions[i].score(docId, subQueryScore), minFactor);
                        }
                    }
                    if (minFactor != Float.POSITIVE_INFINITY) {
                        factor = minFactor;
                    }
                    break;
                case MULTIPLY:
                    for (int i = 0; i < filterFunctions.length; i++) {
                        if (docSets[i].get(docId)) {
                            factor *= functions[i].score(docId, subQueryScore);
                        }
                    }
                    break;
                default: // Avg / Total
                    double totalFactor = 0.0f;
                    double weightSum = 0;
                    for (int i = 0; i < filterFunctions.length; i++) {
                        if (docSets[i].get(docId)) {
                            totalFactor += functions[i].score(docId, subQueryScore);
                            if (filterFunctions[i].function instanceof WeightFactorFunction) {
                                weightSum += ((WeightFactorFunction) filterFunctions[i].function).getWeight();
                            } else {
                                weightSum += 1.0;
                            }
                        }
                    }
                    if (weightSum != 0) {
                        factor = totalFactor;
                        if (scoreMode == ScoreMode.AVG) {
                            factor /= weightSum;
                        }
                    }
                    break;
            }
            return factor;
        }
    }

    @Override
    public String toString(String field) {
        StringBuilder sb = new StringBuilder();
        sb.append("function score (").append(subQuery.toString(field)).append(", functions: [");
        for (FilterFunction filterFunction : filterFunctions) {
            sb.append("{filter(").append(filterFunction.filter).append("), function [").append(filterFunction.function).append("]}");
        }
        sb.append("])");
        return sb.toString();
    }

    @Override
    public boolean equals(Object o) {
        if (this == o) {
            return true;
        }
        if (sameClassAs(o) == false) {
            return false;
        }
        FiltersFunctionScoreQuery other = (FiltersFunctionScoreQuery) o;
        return Objects.equals(this.subQuery, other.subQuery) && this.maxBoost == other.maxBoost &&
            Objects.equals(this.combineFunction, other.combineFunction) && Objects.equals(this.minScore, other.minScore) &&
            Objects.equals(this.scoreMode, other.scoreMode) &&
            Arrays.equals(this.filterFunctions, other.filterFunctions);
    }

    @Override
    public int hashCode() {
        return Objects.hash(classHash(), subQuery, maxBoost, combineFunction, minScore, scoreMode, Arrays.hashCode(filterFunctions));
    }
}




© 2015 - 2024 Weber Informatics LLC | Privacy Policy