honeymoose
/
OpenSearch
mirror of https://github.com/honeymoose/OpenSearch.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

percolator: Take `matchAllDocs` and `verified` of the sub result into account when analyzing a function_score query. 

Before the `matchAllDocs` was ignored and this could lead to percolator queries not matching when
the inner query was a match_all query and min_score was specified.

Before when `verified` was not taken into account if the function_score query wrapped an unverified query this could
lead to matching percolator queries that shouldn't match at all.
This commit is contained in:
Martijn van Groningen 2018-02-28 10:06:47 -08:00
parent 0f95636a91
commit beb22d89c8
3 changed files with 62 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
modules/percolator/src/main/java/org/elasticsearch/percolator/QueryAnalyzer.java
View File

@ -465,12 +465,17 @@ final class QueryAnalyzer {
return (query, version) -> {
FunctionScoreQuery functionScoreQuery = (FunctionScoreQuery) query;
Result result = analyze(functionScoreQuery.getSubQuery(), version);
// If min_score is specified we can't guarantee upfront that this percolator query matches,
// so in that case we set verified to false.
// (if it matches with the percolator document matches with the extracted terms.
// Min score filters out docs, which is different than the functions, which just influences the score.)
boolean verified = functionScoreQuery.getMinScore() == null;
boolean verified = result.verified && functionScoreQuery.getMinScore() == null;
if (result.matchAllDocs) {
return new Result(result.matchAllDocs, verified);
} else {
return new Result(verified, result.extractions, result.minimumShouldMatch);
}
};
}

37
modules/percolator/src/test/java/org/elasticsearch/percolator/CandidateQueryTests.java
View File

@ -55,6 +55,7 @@ import org.apache.lucene.search.FilteredDocIdSetIterator;
import org.apache.lucene.search.IndexSearcher;
import org.apache.lucene.search.MatchAllDocsQuery;
import org.apache.lucene.search.MatchNoDocsQuery;
import org.apache.lucene.search.PhraseQuery;
import org.apache.lucene.search.PrefixQuery;
import org.apache.lucene.search.Query;
import org.apache.lucene.search.Scorer;
@ -77,6 +78,7 @@ import org.elasticsearch.common.CheckedFunction;
import org.elasticsearch.common.bytes.BytesArray;
import org.elasticsearch.common.bytes.BytesReference;
import org.elasticsearch.common.compress.CompressedXContent;
import org.elasticsearch.common.lucene.search.function.FunctionScoreQuery;
import org.elasticsearch.common.geo.ShapeRelation;
import org.elasticsearch.common.settings.Settings;
import org.elasticsearch.common.xcontent.XContentFactory;
@ -220,6 +222,16 @@ public class CandidateQueryTests extends ESSingleNodeTestCase {
}
return new DisjunctionMaxQuery(clauses, 0.01f);
});
queryFunctions.add(() -> {
Float minScore = randomBoolean() ? null : (float) randomIntBetween(1, 1000);
Query innerQuery;
if (randomBoolean()) {
innerQuery = new TermQuery(new Term(field1, randomFrom(stringContent.get(field1))));
} else {
innerQuery = new PhraseQuery(field1, randomFrom(stringContent.get(field1)), randomFrom(stringContent.get(field1)));
}
return new FunctionScoreQuery(innerQuery, minScore, 1f);
});
List<ParseContext.Document> documents = new ArrayList<>();
for (Supplier<Query> queryFunction : queryFunctions) {
@ -679,6 +691,31 @@ public class CandidateQueryTests extends ESSingleNodeTestCase {
assertEquals(4, topDocs.scoreDocs[2].doc);
}
public void testFunctionScoreQuery() throws Exception {
List<ParseContext.Document> docs = new ArrayList<>();
addQuery(new FunctionScoreQuery(new TermQuery(new Term("field", "value")), null, 1f), docs);
addQuery(new FunctionScoreQuery(new TermQuery(new Term("field", "value")), 10f, 1f), docs);
addQuery(new FunctionScoreQuery(new MatchAllDocsQuery(), null, 1f), docs);
addQuery(new FunctionScoreQuery(new MatchAllDocsQuery(), 10F, 1f), docs);
indexWriter.addDocuments(docs);
indexWriter.close();
directoryReader = DirectoryReader.open(directory);
IndexSearcher shardSearcher = newSearcher(directoryReader);
shardSearcher.setQueryCache(null);
MemoryIndex memoryIndex = new MemoryIndex();
memoryIndex.addField("field", "value", new WhitespaceAnalyzer());
IndexSearcher percolateSearcher = memoryIndex.createSearcher();
PercolateQuery query = (PercolateQuery) fieldType.percolateQuery("_name", queryStore,
Collections.singletonList(new BytesArray("{}")), percolateSearcher, Version.CURRENT);
TopDocs topDocs = shardSearcher.search(query, 10, new Sort(SortField.FIELD_DOC), true, true);
assertEquals(2L, topDocs.totalHits);
assertEquals(2, topDocs.scoreDocs.length);
assertEquals(0, topDocs.scoreDocs[0].doc);
assertEquals(2, topDocs.scoreDocs[1].doc);
}
public void testPercolateSmallAndLargeDocument() throws Exception {
List<ParseContext.Document> docs = new ArrayList<>();
BooleanQuery.Builder builder = new BooleanQuery.Builder();

18
modules/percolator/src/test/java/org/elasticsearch/percolator/QueryAnalyzerTests.java
View File

@ -811,6 +811,24 @@ public class QueryAnalyzerTests extends ESTestCase {
assertTermsEqual(result.extractions, new Term("_field", "_value"));
}
public void testFunctionScoreQuery_withMatchAll() {
MatchAllDocsQuery innerQuery = new MatchAllDocsQuery();
FunctionScoreQuery functionScoreQuery1 = new FunctionScoreQuery(innerQuery, new RandomScoreFunction(0, 0, null));
Result result = analyze(functionScoreQuery1, Version.CURRENT);
assertThat(result.verified, is(true));
assertThat(result.minimumShouldMatch, equalTo(0));
assertThat(result.matchAllDocs, is(true));
assertThat(result.extractions.isEmpty(), is(true));
FunctionScoreQuery functionScoreQuery2 =
new FunctionScoreQuery(innerQuery, new RandomScoreFunction(0, 0, null), CombineFunction.MULTIPLY, 1f, 10f);
result = analyze(functionScoreQuery2, Version.CURRENT);
assertThat(result.verified, is(false));
assertThat(result.minimumShouldMatch, equalTo(0));
assertThat(result.matchAllDocs, is(true));
assertThat(result.extractions.isEmpty(), is(true));
}
public void testSelectBestExtraction() {
Set<QueryExtraction> queryTerms1 = terms(new int[0], "12", "1234", "12345");
Set<QueryAnalyzer.QueryExtraction> queryTerms2 = terms(new int[0], "123", "1234", "12345");