use of org.apache.lucene.search.IndexSearcher in project elasticsearch by elastic.
the class QueryProfilerTests method setup.
@BeforeClass
public static void setup() throws IOException {
dir = newDirectory();
RandomIndexWriter w = new RandomIndexWriter(random(), dir);
final int numDocs = TestUtil.nextInt(random(), 1, 20);
for (int i = 0; i < numDocs; ++i) {
final int numHoles = random().nextInt(5);
for (int j = 0; j < numHoles; ++j) {
w.addDocument(new Document());
}
Document doc = new Document();
doc.add(new StringField("foo", "bar", Store.NO));
w.addDocument(doc);
}
reader = w.getReader();
w.close();
Engine.Searcher engineSearcher = new Engine.Searcher("test", new IndexSearcher(reader));
searcher = new ContextIndexSearcher(engineSearcher, IndexSearcher.getDefaultQueryCache(), MAYBE_CACHE_POLICY);
}
use of org.apache.lucene.search.IndexSearcher in project elasticsearch by elastic.
the class PercolateQueryBuilder method doToQuery.
@Override
protected Query doToQuery(QueryShardContext context) throws IOException {
// Call nowInMillis() so that this query becomes un-cacheable since we
// can't be sure that it doesn't use now or scripts
context.nowInMillis();
if (indexedDocumentIndex != null || indexedDocumentType != null || indexedDocumentId != null) {
throw new IllegalStateException("query builder must be rewritten first");
}
if (document == null) {
throw new IllegalStateException("no document to percolate");
}
MapperService mapperService = context.getMapperService();
DocumentMapperForType docMapperForType = mapperService.documentMapperWithAutoCreate(documentType);
DocumentMapper docMapper = docMapperForType.getDocumentMapper();
ParsedDocument doc = docMapper.parse(source(context.index().getName(), documentType, "_temp_id", document, documentXContentType));
FieldNameAnalyzer fieldNameAnalyzer = (FieldNameAnalyzer) docMapper.mappers().indexAnalyzer();
// Need to this custom impl because FieldNameAnalyzer is strict and the percolator sometimes isn't when
// 'index.percolator.map_unmapped_fields_as_string' is enabled:
Analyzer analyzer = new DelegatingAnalyzerWrapper(Analyzer.PER_FIELD_REUSE_STRATEGY) {
@Override
protected Analyzer getWrappedAnalyzer(String fieldName) {
Analyzer analyzer = fieldNameAnalyzer.analyzers().get(fieldName);
if (analyzer != null) {
return analyzer;
} else {
return context.getIndexAnalyzers().getDefaultIndexAnalyzer();
}
}
};
final IndexSearcher docSearcher;
if (doc.docs().size() > 1) {
assert docMapper.hasNestedObjects();
docSearcher = createMultiDocumentSearcher(analyzer, doc);
} else {
MemoryIndex memoryIndex = MemoryIndex.fromDocument(doc.rootDoc(), analyzer, true, false);
docSearcher = memoryIndex.createSearcher();
docSearcher.setQueryCache(null);
}
Version indexVersionCreated = context.getIndexSettings().getIndexVersionCreated();
boolean mapUnmappedFieldsAsString = context.getIndexSettings().getValue(PercolatorFieldMapper.INDEX_MAP_UNMAPPED_FIELDS_AS_STRING_SETTING);
// We have to make a copy of the QueryShardContext here so we can have a unfrozen version for parsing the legacy
// percolator queries
QueryShardContext percolateShardContext = new QueryShardContext(context);
MappedFieldType fieldType = context.fieldMapper(field);
if (fieldType == null) {
throw new QueryShardException(context, "field [" + field + "] does not exist");
}
if (!(fieldType instanceof PercolatorFieldMapper.FieldType)) {
throw new QueryShardException(context, "expected field [" + field + "] to be of type [percolator], but is of type [" + fieldType.typeName() + "]");
}
PercolatorFieldMapper.FieldType pft = (PercolatorFieldMapper.FieldType) fieldType;
PercolateQuery.QueryStore queryStore = createStore(pft, percolateShardContext, mapUnmappedFieldsAsString);
return pft.percolateQuery(documentType, queryStore, document, docSearcher);
}
use of org.apache.lucene.search.IndexSearcher in project elasticsearch by elastic.
the class PercolateQueryBuilder method createMultiDocumentSearcher.
static IndexSearcher createMultiDocumentSearcher(Analyzer analyzer, ParsedDocument doc) {
RAMDirectory ramDirectory = new RAMDirectory();
try (IndexWriter indexWriter = new IndexWriter(ramDirectory, new IndexWriterConfig(analyzer))) {
indexWriter.addDocuments(doc.docs());
indexWriter.commit();
DirectoryReader directoryReader = DirectoryReader.open(ramDirectory);
assert directoryReader.leaves().size() == 1 : "Expected single leaf, but got [" + directoryReader.leaves().size() + "]";
final IndexSearcher slowSearcher = new IndexSearcher(directoryReader) {
@Override
public Weight createNormalizedWeight(Query query, boolean needsScores) throws IOException {
BooleanQuery.Builder bq = new BooleanQuery.Builder();
bq.add(query, BooleanClause.Occur.MUST);
bq.add(Queries.newNestedFilter(), BooleanClause.Occur.MUST_NOT);
return super.createNormalizedWeight(bq.build(), needsScores);
}
};
slowSearcher.setQueryCache(null);
return slowSearcher;
} catch (IOException e) {
throw new ElasticsearchException("Failed to create index for percolator with nested document ", e);
}
}
use of org.apache.lucene.search.IndexSearcher in project elasticsearch by elastic.
the class CandidateQueryTests method testDuel.
public void testDuel() throws Exception {
List<Function<String, Query>> queryFunctions = new ArrayList<>();
queryFunctions.add((id) -> new PrefixQuery(new Term("field", id)));
queryFunctions.add((id) -> new WildcardQuery(new Term("field", id + "*")));
queryFunctions.add((id) -> new CustomQuery(new Term("field", id)));
queryFunctions.add((id) -> new SpanTermQuery(new Term("field", id)));
queryFunctions.add((id) -> new TermQuery(new Term("field", id)));
queryFunctions.add((id) -> {
BooleanQuery.Builder builder = new BooleanQuery.Builder();
return builder.build();
});
queryFunctions.add((id) -> {
BooleanQuery.Builder builder = new BooleanQuery.Builder();
builder.add(new TermQuery(new Term("field", id)), BooleanClause.Occur.MUST);
if (randomBoolean()) {
builder.add(new MatchNoDocsQuery("no reason"), BooleanClause.Occur.MUST_NOT);
}
if (randomBoolean()) {
builder.add(new CustomQuery(new Term("field", id)), BooleanClause.Occur.MUST);
}
return builder.build();
});
queryFunctions.add((id) -> {
BooleanQuery.Builder builder = new BooleanQuery.Builder();
builder.add(new TermQuery(new Term("field", id)), BooleanClause.Occur.SHOULD);
if (randomBoolean()) {
builder.add(new MatchNoDocsQuery("no reason"), BooleanClause.Occur.MUST_NOT);
}
if (randomBoolean()) {
builder.add(new CustomQuery(new Term("field", id)), BooleanClause.Occur.SHOULD);
}
return builder.build();
});
queryFunctions.add((id) -> {
BooleanQuery.Builder builder = new BooleanQuery.Builder();
builder.add(new MatchAllDocsQuery(), BooleanClause.Occur.MUST);
builder.add(new MatchAllDocsQuery(), BooleanClause.Occur.MUST);
if (randomBoolean()) {
builder.add(new MatchNoDocsQuery("no reason"), BooleanClause.Occur.MUST_NOT);
}
return builder.build();
});
queryFunctions.add((id) -> {
BooleanQuery.Builder builder = new BooleanQuery.Builder();
builder.add(new MatchAllDocsQuery(), BooleanClause.Occur.SHOULD);
builder.add(new MatchAllDocsQuery(), BooleanClause.Occur.SHOULD);
if (randomBoolean()) {
builder.add(new MatchNoDocsQuery("no reason"), BooleanClause.Occur.MUST_NOT);
}
return builder.build();
});
queryFunctions.add((id) -> {
BooleanQuery.Builder builder = new BooleanQuery.Builder();
builder.setMinimumNumberShouldMatch(randomIntBetween(0, 4));
builder.add(new TermQuery(new Term("field", id)), BooleanClause.Occur.SHOULD);
builder.add(new CustomQuery(new Term("field", id)), BooleanClause.Occur.SHOULD);
return builder.build();
});
queryFunctions.add((id) -> new MatchAllDocsQuery());
queryFunctions.add((id) -> new MatchNoDocsQuery("no reason at all"));
int numDocs = randomIntBetween(queryFunctions.size(), queryFunctions.size() * 3);
List<ParseContext.Document> documents = new ArrayList<>();
for (int i = 0; i < numDocs; i++) {
String id = Integer.toString(i);
Query query = queryFunctions.get(i % queryFunctions.size()).apply(id);
addQuery(query, documents);
}
indexWriter.addDocuments(documents);
indexWriter.close();
directoryReader = DirectoryReader.open(directory);
IndexSearcher shardSearcher = newSearcher(directoryReader);
// Disable query cache, because ControlQuery cannot be cached...
shardSearcher.setQueryCache(null);
for (int i = 0; i < numDocs; i++) {
String id = Integer.toString(i);
Iterable<? extends IndexableField> doc = Collections.singleton(new StringField("field", id, Field.Store.NO));
MemoryIndex memoryIndex = MemoryIndex.fromDocument(doc, new WhitespaceAnalyzer());
duelRun(queryStore, memoryIndex, shardSearcher);
}
Iterable<? extends IndexableField> doc = Collections.singleton(new StringField("field", "value", Field.Store.NO));
MemoryIndex memoryIndex = MemoryIndex.fromDocument(doc, new WhitespaceAnalyzer());
duelRun(queryStore, memoryIndex, shardSearcher);
// Empty percolator doc:
memoryIndex = new MemoryIndex();
duelRun(queryStore, memoryIndex, shardSearcher);
}
use of org.apache.lucene.search.IndexSearcher in project elasticsearch by elastic.
the class PercolateQueryBuilderTests method testCreateMultiDocumentSearcher.
public void testCreateMultiDocumentSearcher() throws Exception {
int numDocs = randomIntBetween(2, 8);
List<ParseContext.Document> docs = new ArrayList<>(numDocs);
for (int i = 0; i < numDocs; i++) {
docs.add(new ParseContext.Document());
}
Analyzer analyzer = new WhitespaceAnalyzer();
ParsedDocument parsedDocument = new ParsedDocument(null, null, "_id", "_type", null, docs, null, null, null);
IndexSearcher indexSearcher = PercolateQueryBuilder.createMultiDocumentSearcher(analyzer, parsedDocument);
assertThat(indexSearcher.getIndexReader().numDocs(), equalTo(numDocs));
// ensure that any query get modified so that the nested docs are never included as hits:
Query query = new MatchAllDocsQuery();
BooleanQuery result = (BooleanQuery) indexSearcher.createNormalizedWeight(query, true).getQuery();
assertThat(result.clauses().size(), equalTo(2));
assertThat(result.clauses().get(0).getQuery(), sameInstance(query));
assertThat(result.clauses().get(0).getOccur(), equalTo(BooleanClause.Occur.MUST));
assertThat(result.clauses().get(1).getOccur(), equalTo(BooleanClause.Occur.MUST_NOT));
}
Aggregations