mirror of https://github.com/apache/lucene.git
LUCENE-6844: Add includeSpanScore option to PayloadScoreQuery
git-svn-id: https://svn.apache.org/repos/asf/lucene/dev/trunk@1709386 13f79535-47bb-0310-9956-ffa450edef68
This commit is contained in:
parent
c1ec5f5ec7
commit
0b305f9279
|
@ -91,6 +91,9 @@ New Features
|
||||||
* LUCENE-6838: Added IndexSearcher#getQueryCache and #getQueryCachingPolicy.
|
* LUCENE-6838: Added IndexSearcher#getQueryCache and #getQueryCachingPolicy.
|
||||||
(Adrien Grand)
|
(Adrien Grand)
|
||||||
|
|
||||||
|
* LUCENE-6844: PayloadScoreQuery can include or exclude underlying span scores
|
||||||
|
from its score calculations (Bill Bell, Alan Woodward)
|
||||||
|
|
||||||
API Changes
|
API Changes
|
||||||
|
|
||||||
* LUCENE-6590: Query.setBoost(), Query.getBoost() and Query.clone() are gone.
|
* LUCENE-6590: Query.setBoost(), Query.getBoost() and Query.clone() are gone.
|
||||||
|
|
|
@ -51,15 +51,27 @@ public class PayloadScoreQuery extends SpanQuery {
|
||||||
|
|
||||||
private final SpanQuery wrappedQuery;
|
private final SpanQuery wrappedQuery;
|
||||||
private final PayloadFunction function;
|
private final PayloadFunction function;
|
||||||
|
private final boolean includeSpanScore;
|
||||||
|
|
||||||
/**
|
/**
|
||||||
* Creates a new PayloadScoreQuery
|
* Creates a new PayloadScoreQuery
|
||||||
* @param wrappedQuery the query to wrap
|
* @param wrappedQuery the query to wrap
|
||||||
* @param function a PayloadFunction to use to modify the scores
|
* @param function a PayloadFunction to use to modify the scores
|
||||||
|
* @param includeSpanScore include both span score and payload score in the scoring algorithm
|
||||||
*/
|
*/
|
||||||
public PayloadScoreQuery(SpanQuery wrappedQuery, PayloadFunction function) {
|
public PayloadScoreQuery(SpanQuery wrappedQuery, PayloadFunction function, boolean includeSpanScore) {
|
||||||
this.wrappedQuery = wrappedQuery;
|
this.wrappedQuery = wrappedQuery;
|
||||||
this.function = function;
|
this.function = function;
|
||||||
|
this.includeSpanScore = includeSpanScore;
|
||||||
|
}
|
||||||
|
|
||||||
|
/**
|
||||||
|
* Creates a new PayloadScoreQuery that includes the underlying span scores
|
||||||
|
* @param wrappedQuery the query to wrap
|
||||||
|
* @param function a PayloadFunction to use to modify the scores
|
||||||
|
*/
|
||||||
|
public PayloadScoreQuery(SpanQuery wrappedQuery, PayloadFunction function) {
|
||||||
|
this(wrappedQuery, function, true);
|
||||||
}
|
}
|
||||||
|
|
||||||
@Override
|
@Override
|
||||||
|
@ -149,12 +161,16 @@ public class PayloadScoreQuery extends SpanQuery {
|
||||||
if (scorer == null || scorer.advance(doc) != doc)
|
if (scorer == null || scorer.advance(doc) != doc)
|
||||||
return Explanation.noMatch("No match");
|
return Explanation.noMatch("No match");
|
||||||
|
|
||||||
SpanWeight innerWeight = ((PayloadSpanWeight)scorer.getWeight()).innerWeight;
|
|
||||||
Explanation innerExpl = innerWeight.explain(context, doc);
|
|
||||||
scorer.freq(); // force freq calculation
|
scorer.freq(); // force freq calculation
|
||||||
Explanation payloadExpl = scorer.getPayloadExplanation();
|
Explanation payloadExpl = scorer.getPayloadExplanation();
|
||||||
|
|
||||||
return Explanation.match(scorer.scoreCurrentDoc(), "PayloadSpanQuery, product of:", innerExpl, payloadExpl);
|
if (includeSpanScore) {
|
||||||
|
SpanWeight innerWeight = ((PayloadSpanWeight) scorer.getWeight()).innerWeight;
|
||||||
|
Explanation innerExpl = innerWeight.explain(context, doc);
|
||||||
|
return Explanation.match(scorer.scoreCurrentDoc(), "PayloadSpanQuery, product of:", innerExpl, payloadExpl);
|
||||||
|
}
|
||||||
|
|
||||||
|
return scorer.getPayloadExplanation();
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
|
@ -203,7 +219,9 @@ public class PayloadScoreQuery extends SpanQuery {
|
||||||
|
|
||||||
@Override
|
@Override
|
||||||
protected float scoreCurrentDoc() throws IOException {
|
protected float scoreCurrentDoc() throws IOException {
|
||||||
return getSpanScore() * getPayloadScore();
|
if (includeSpanScore)
|
||||||
|
return getSpanScore() * getPayloadScore();
|
||||||
|
return getPayloadScore();
|
||||||
}
|
}
|
||||||
|
|
||||||
@Override
|
@Override
|
||||||
|
|
|
@ -51,7 +51,7 @@ public class TestPayloadExplanations extends BaseExplanationTestCase {
|
||||||
|
|
||||||
/** macro for payloadscorequery */
|
/** macro for payloadscorequery */
|
||||||
private SpanQuery pt(String s, PayloadFunction fn) {
|
private SpanQuery pt(String s, PayloadFunction fn) {
|
||||||
return new PayloadScoreQuery(new SpanTermQuery(new Term(FIELD,s)), fn);
|
return new PayloadScoreQuery(new SpanTermQuery(new Term(FIELD,s)), fn, random().nextBoolean());
|
||||||
}
|
}
|
||||||
|
|
||||||
/* simple PayloadTermQueries */
|
/* simple PayloadTermQueries */
|
||||||
|
@ -82,8 +82,6 @@ public class TestPayloadExplanations extends BaseExplanationTestCase {
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
// TODO: test the payloadnear query too!
|
|
||||||
|
|
||||||
/*
|
/*
|
||||||
protected static final String[] docFields = {
|
protected static final String[] docFields = {
|
||||||
"w1 w2 w3 w4 w5",
|
"w1 w2 w3 w4 w5",
|
||||||
|
@ -95,7 +93,7 @@ public class TestPayloadExplanations extends BaseExplanationTestCase {
|
||||||
|
|
||||||
public void testAllFunctions(SpanQuery query, int[] expected) throws Exception {
|
public void testAllFunctions(SpanQuery query, int[] expected) throws Exception {
|
||||||
for (PayloadFunction fn : functions) {
|
for (PayloadFunction fn : functions) {
|
||||||
qtest(new PayloadScoreQuery(query, fn), expected);
|
qtest(new PayloadScoreQuery(query, fn, random().nextBoolean()), expected);
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
|
|
|
@ -35,6 +35,7 @@ import org.apache.lucene.index.Term;
|
||||||
import org.apache.lucene.search.CollectionStatistics;
|
import org.apache.lucene.search.CollectionStatistics;
|
||||||
import org.apache.lucene.search.Explanation;
|
import org.apache.lucene.search.Explanation;
|
||||||
import org.apache.lucene.search.IndexSearcher;
|
import org.apache.lucene.search.IndexSearcher;
|
||||||
|
import org.apache.lucene.search.QueryUtils;
|
||||||
import org.apache.lucene.search.TermStatistics;
|
import org.apache.lucene.search.TermStatistics;
|
||||||
import org.apache.lucene.search.TopDocs;
|
import org.apache.lucene.search.TopDocs;
|
||||||
import org.apache.lucene.search.similarities.ClassicSimilarity;
|
import org.apache.lucene.search.similarities.ClassicSimilarity;
|
||||||
|
@ -54,10 +55,14 @@ import org.junit.Test;
|
||||||
public class TestPayloadScoreQuery extends LuceneTestCase {
|
public class TestPayloadScoreQuery extends LuceneTestCase {
|
||||||
|
|
||||||
private static void checkQuery(SpanQuery query, PayloadFunction function, int[] expectedDocs, float[] expectedScores) throws IOException {
|
private static void checkQuery(SpanQuery query, PayloadFunction function, int[] expectedDocs, float[] expectedScores) throws IOException {
|
||||||
|
checkQuery(query, function, true, expectedDocs, expectedScores);
|
||||||
|
}
|
||||||
|
|
||||||
|
private static void checkQuery(SpanQuery query, PayloadFunction function, boolean includeSpanScore, int[] expectedDocs, float[] expectedScores) throws IOException {
|
||||||
|
|
||||||
assertTrue("Expected docs and scores arrays must be the same length!", expectedDocs.length == expectedScores.length);
|
assertTrue("Expected docs and scores arrays must be the same length!", expectedDocs.length == expectedScores.length);
|
||||||
|
|
||||||
PayloadScoreQuery psq = new PayloadScoreQuery(query, function);
|
PayloadScoreQuery psq = new PayloadScoreQuery(query, function, includeSpanScore);
|
||||||
TopDocs hits = searcher.search(psq, expectedDocs.length);
|
TopDocs hits = searcher.search(psq, expectedDocs.length);
|
||||||
|
|
||||||
for (int i = 0; i < hits.scoreDocs.length; i++) {
|
for (int i = 0; i < hits.scoreDocs.length; i++) {
|
||||||
|
@ -70,6 +75,8 @@ public class TestPayloadScoreQuery extends LuceneTestCase {
|
||||||
|
|
||||||
if (hits.scoreDocs.length > expectedDocs.length)
|
if (hits.scoreDocs.length > expectedDocs.length)
|
||||||
fail("Unexpected hit in document " + hits.scoreDocs[expectedDocs.length]);
|
fail("Unexpected hit in document " + hits.scoreDocs[expectedDocs.length]);
|
||||||
|
|
||||||
|
QueryUtils.check(random(), psq, searcher);
|
||||||
}
|
}
|
||||||
|
|
||||||
@Test
|
@Test
|
||||||
|
@ -132,9 +139,19 @@ public class TestPayloadScoreQuery extends LuceneTestCase {
|
||||||
}, 0, true)
|
}, 0, true)
|
||||||
}, 1, true);
|
}, 1, true);
|
||||||
|
|
||||||
checkQuery(q, new MaxPayloadFunction(), new int[]{ 122, 222 }, new float[]{ 4.0f, 4.0f });
|
// check includeSpanScore makes a difference here
|
||||||
checkQuery(q, new MinPayloadFunction(), new int[]{ 222, 122 }, new float[]{ 4.0f, 2.0f });
|
searcher.setSimilarity(new MultiplyingSimilarity());
|
||||||
checkQuery(q, new AveragePayloadFunction(), new int[] { 222, 122 }, new float[]{ 4.0f, 3.666666f });
|
try {
|
||||||
|
checkQuery(q, new MaxPayloadFunction(), new int[]{ 122, 222 }, new float[]{ 41.802513122558594f, 34.13160705566406f });
|
||||||
|
checkQuery(q, new MinPayloadFunction(), new int[]{ 222, 122 }, new float[]{ 34.13160705566406f, 20.901256561279297f });
|
||||||
|
checkQuery(q, new AveragePayloadFunction(), new int[] { 122, 222 }, new float[]{ 38.3189697265625f, 34.13160705566406f });
|
||||||
|
checkQuery(q, new MaxPayloadFunction(), false, new int[]{122, 222}, new float[]{4.0f, 4.0f});
|
||||||
|
checkQuery(q, new MinPayloadFunction(), false, new int[]{222, 122}, new float[]{4.0f, 2.0f});
|
||||||
|
checkQuery(q, new AveragePayloadFunction(), false, new int[]{222, 122}, new float[]{4.0f, 3.666666f});
|
||||||
|
}
|
||||||
|
finally {
|
||||||
|
searcher.setSimilarity(similarity);
|
||||||
|
}
|
||||||
|
|
||||||
}
|
}
|
||||||
|
|
||||||
|
@ -234,7 +251,17 @@ public class TestPayloadScoreQuery extends LuceneTestCase {
|
||||||
directory = null;
|
directory = null;
|
||||||
}
|
}
|
||||||
|
|
||||||
static class BoostingSimilarity extends ClassicSimilarity {
|
static class MultiplyingSimilarity extends ClassicSimilarity {
|
||||||
|
|
||||||
|
@Override
|
||||||
|
public float scorePayload(int docId, int start, int end, BytesRef payload) {
|
||||||
|
//we know it is size 4 here, so ignore the offset/length
|
||||||
|
return payload.bytes[payload.offset];
|
||||||
|
}
|
||||||
|
|
||||||
|
}
|
||||||
|
|
||||||
|
static class BoostingSimilarity extends MultiplyingSimilarity {
|
||||||
|
|
||||||
@Override
|
@Override
|
||||||
public float queryNorm(float sumOfSquaredWeights) {
|
public float queryNorm(float sumOfSquaredWeights) {
|
||||||
|
@ -246,12 +273,6 @@ public class TestPayloadScoreQuery extends LuceneTestCase {
|
||||||
return 1.0f;
|
return 1.0f;
|
||||||
}
|
}
|
||||||
|
|
||||||
@Override
|
|
||||||
public float scorePayload(int docId, int start, int end, BytesRef payload) {
|
|
||||||
//we know it is size 4 here, so ignore the offset/length
|
|
||||||
return payload.bytes[payload.offset];
|
|
||||||
}
|
|
||||||
|
|
||||||
//!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
|
//!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
|
||||||
//Make everything else 1 so we see the effect of the payload
|
//Make everything else 1 so we see the effect of the payload
|
||||||
//!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
|
//!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
|
||||||
|
|
Loading…
Reference in New Issue