@Test
public void testNestedChildrenFilter() throws Exception {
int numParentDocs = scaledRandomIntBetween(0, 32);
int maxChildDocsPerParent = scaledRandomIntBetween(8, 16);
Directory dir = newDirectory();
RandomIndexWriter writer = new RandomIndexWriter(random(), dir);
for (int i = 0; i < numParentDocs; i++) {
int numChildDocs = scaledRandomIntBetween(0, maxChildDocsPerParent);
List<Document> docs = new ArrayList<>(numChildDocs + 1);
for (int j = 0; j < numChildDocs; j++) {
Document childDoc = new Document();
childDoc.add(new StringField("type", "child", Field.Store.NO));
docs.add(childDoc);
}
Document parenDoc = new Document();
parenDoc.add(new StringField("type", "parent", Field.Store.NO));
parenDoc.add(new IntField("num_child_docs", numChildDocs, Field.Store.YES));
docs.add(parenDoc);
writer.addDocuments(docs);
}
IndexReader reader = writer.getReader();
writer.close();
IndexSearcher searcher = new IndexSearcher(reader);
FetchSubPhase.HitContext hitContext = new FetchSubPhase.HitContext();
BitDocIdSetFilter parentFilter =
new BitDocIdSetCachingWrapperFilter(
new QueryWrapperFilter(new TermQuery(new Term("type", "parent"))));
Filter childFilter = new QueryWrapperFilter(new TermQuery(new Term("type", "child")));
int checkedParents = 0;
for (LeafReaderContext leaf : reader.leaves()) {
DocIdSetIterator parents = parentFilter.getDocIdSet(leaf).iterator();
for (int parentDoc = parents.nextDoc();
parentDoc != DocIdSetIterator.NO_MORE_DOCS;
parentDoc = parents.nextDoc()) {
int expectedChildDocs =
leaf.reader().document(parentDoc).getField("num_child_docs").numericValue().intValue();
hitContext.reset(null, leaf, parentDoc, searcher);
NestedChildrenFilter nestedChildrenFilter =
new NestedChildrenFilter(parentFilter, childFilter, hitContext);
TotalHitCountCollector totalHitCountCollector = new TotalHitCountCollector();
searcher.search(new ConstantScoreQuery(nestedChildrenFilter), totalHitCountCollector);
assertThat(totalHitCountCollector.getTotalHits(), equalTo(expectedChildDocs));
checkedParents++;
}
}
assertThat(checkedParents, equalTo(numParentDocs));
reader.close();
dir.close();
}
public void testResetRootDocId() throws Exception {
Directory directory = newDirectory();
IndexWriterConfig iwc = new IndexWriterConfig(null);
iwc.setMergePolicy(NoMergePolicy.INSTANCE);
RandomIndexWriter indexWriter = new RandomIndexWriter(random(), directory, iwc);
List<Document> documents = new ArrayList<>();
// 1 segment with, 1 root document, with 3 nested sub docs
Document document = new Document();
document.add(
new Field(UidFieldMapper.NAME, "type#1", UidFieldMapper.Defaults.NESTED_FIELD_TYPE));
document.add(
new Field(TypeFieldMapper.NAME, "__nested_field", TypeFieldMapper.Defaults.FIELD_TYPE));
documents.add(document);
document = new Document();
document.add(
new Field(UidFieldMapper.NAME, "type#1", UidFieldMapper.Defaults.NESTED_FIELD_TYPE));
document.add(
new Field(TypeFieldMapper.NAME, "__nested_field", TypeFieldMapper.Defaults.FIELD_TYPE));
documents.add(document);
document = new Document();
document.add(
new Field(UidFieldMapper.NAME, "type#1", UidFieldMapper.Defaults.NESTED_FIELD_TYPE));
document.add(
new Field(TypeFieldMapper.NAME, "__nested_field", TypeFieldMapper.Defaults.FIELD_TYPE));
documents.add(document);
document = new Document();
document.add(new Field(UidFieldMapper.NAME, "type#1", UidFieldMapper.Defaults.FIELD_TYPE));
document.add(new Field(TypeFieldMapper.NAME, "test", TypeFieldMapper.Defaults.FIELD_TYPE));
documents.add(document);
indexWriter.addDocuments(documents);
indexWriter.commit();
documents.clear();
// 1 segment with:
// 1 document, with 1 nested subdoc
document = new Document();
document.add(
new Field(UidFieldMapper.NAME, "type#2", UidFieldMapper.Defaults.NESTED_FIELD_TYPE));
document.add(
new Field(TypeFieldMapper.NAME, "__nested_field", TypeFieldMapper.Defaults.FIELD_TYPE));
documents.add(document);
document = new Document();
document.add(new Field(UidFieldMapper.NAME, "type#2", UidFieldMapper.Defaults.FIELD_TYPE));
document.add(new Field(TypeFieldMapper.NAME, "test", TypeFieldMapper.Defaults.FIELD_TYPE));
documents.add(document);
indexWriter.addDocuments(documents);
documents.clear();
// and 1 document, with 1 nested subdoc
document = new Document();
document.add(
new Field(UidFieldMapper.NAME, "type#3", UidFieldMapper.Defaults.NESTED_FIELD_TYPE));
document.add(
new Field(TypeFieldMapper.NAME, "__nested_field", TypeFieldMapper.Defaults.FIELD_TYPE));
documents.add(document);
document = new Document();
document.add(new Field(UidFieldMapper.NAME, "type#3", UidFieldMapper.Defaults.FIELD_TYPE));
document.add(new Field(TypeFieldMapper.NAME, "test", TypeFieldMapper.Defaults.FIELD_TYPE));
documents.add(document);
indexWriter.addDocuments(documents);
indexWriter.commit();
indexWriter.close();
IndexService indexService = createIndex("test");
DirectoryReader directoryReader = DirectoryReader.open(directory);
directoryReader =
ElasticsearchDirectoryReader.wrap(directoryReader, new ShardId(indexService.index(), 0));
IndexSearcher searcher = new IndexSearcher(directoryReader);
indexService
.mapperService()
.merge(
"test",
new CompressedXContent(
PutMappingRequest.buildFromSimplifiedDef("test", "nested_field", "type=nested")
.string()),
MapperService.MergeReason.MAPPING_UPDATE,
false);
SearchContext searchContext = createSearchContext(indexService);
AggregationContext context = new AggregationContext(searchContext);
AggregatorFactories.Builder builder = AggregatorFactories.builder();
NestedAggregatorBuilder factory = new NestedAggregatorBuilder("test", "nested_field");
builder.addAggregator(factory);
AggregatorFactories factories = builder.build(context, null);
searchContext.aggregations(new SearchContextAggregations(factories));
Aggregator[] aggs = factories.createTopLevelAggregators();
BucketCollector collector = BucketCollector.wrap(Arrays.asList(aggs));
collector.preCollection();
// A regular search always exclude nested docs, so we use NonNestedDocsFilter.INSTANCE here
// (otherwise MatchAllDocsQuery would be sufficient)
// We exclude root doc with uid type#2, this will trigger the bug if we don't reset the root doc
// when we process a new segment, because
// root doc type#3 and root doc type#1 have the same segment docid
BooleanQuery.Builder bq = new BooleanQuery.Builder();
bq.add(Queries.newNonNestedFilter(), Occur.MUST);
bq.add(new TermQuery(new Term(UidFieldMapper.NAME, "type#2")), Occur.MUST_NOT);
searcher.search(new ConstantScoreQuery(bq.build()), collector);
collector.postCollection();
Nested nested = (Nested) aggs[0].buildAggregation(0);
// The bug manifests if 6 docs are returned, because currentRootDoc isn't reset the previous
// child docs from the first segment are emitted as hits.
assertThat(nested.getDocCount(), equalTo(4L));
directoryReader.close();
directory.close();
}
// Tests some very basic usages...
public void testBasic() throws Exception {
final String groupField = "author";
FieldType customType = new FieldType();
customType.setStored(true);
Directory dir = newDirectory();
RandomIndexWriter w =
new RandomIndexWriter(
random(),
dir,
newIndexWriterConfig(new MockAnalyzer(random())).setMergePolicy(newLogMergePolicy()));
boolean canUseIDV = true;
List<Document> documents = new ArrayList<>();
// 0
Document doc = new Document();
addGroupField(doc, groupField, "author1", canUseIDV);
doc.add(new TextField("content", "random text", Field.Store.YES));
doc.add(new Field("id", "1", customType));
documents.add(doc);
// 1
doc = new Document();
addGroupField(doc, groupField, "author1", canUseIDV);
doc.add(new TextField("content", "some more random text", Field.Store.YES));
doc.add(new Field("id", "2", customType));
documents.add(doc);
// 2
doc = new Document();
addGroupField(doc, groupField, "author1", canUseIDV);
doc.add(new TextField("content", "some more random textual data", Field.Store.YES));
doc.add(new Field("id", "3", customType));
doc.add(new StringField("groupend", "x", Field.Store.NO));
documents.add(doc);
w.addDocuments(documents);
documents.clear();
// 3
doc = new Document();
addGroupField(doc, groupField, "author2", canUseIDV);
doc.add(new TextField("content", "some random text", Field.Store.YES));
doc.add(new Field("id", "4", customType));
doc.add(new StringField("groupend", "x", Field.Store.NO));
w.addDocument(doc);
// 4
doc = new Document();
addGroupField(doc, groupField, "author3", canUseIDV);
doc.add(new TextField("content", "some more random text", Field.Store.YES));
doc.add(new Field("id", "5", customType));
documents.add(doc);
// 5
doc = new Document();
addGroupField(doc, groupField, "author3", canUseIDV);
doc.add(new TextField("content", "random", Field.Store.YES));
doc.add(new Field("id", "6", customType));
doc.add(new StringField("groupend", "x", Field.Store.NO));
documents.add(doc);
w.addDocuments(documents);
documents.clear();
// 6 -- no author field
doc = new Document();
doc.add(new TextField("content", "random word stuck in alot of other text", Field.Store.YES));
doc.add(new Field("id", "6", customType));
doc.add(new StringField("groupend", "x", Field.Store.NO));
w.addDocument(doc);
IndexSearcher indexSearcher = newSearcher(w.getReader());
w.close();
Sort groupSort = Sort.RELEVANCE;
GroupingSearch groupingSearch = createRandomGroupingSearch(groupField, groupSort, 5, canUseIDV);
TopGroups<?> groups =
groupingSearch.search(indexSearcher, new TermQuery(new Term("content", "random")), 0, 10);
assertEquals(7, groups.totalHitCount);
assertEquals(7, groups.totalGroupedHitCount);
assertEquals(4, groups.groups.length);
// relevance order: 5, 0, 3, 4, 1, 2, 6
// the later a document is added the higher this docId
// value
GroupDocs<?> group = groups.groups[0];
compareGroupValue("author3", group);
assertEquals(2, group.scoreDocs.length);
assertEquals(5, group.scoreDocs[0].doc);
assertEquals(4, group.scoreDocs[1].doc);
assertTrue(group.scoreDocs[0].score > group.scoreDocs[1].score);
group = groups.groups[1];
compareGroupValue("author1", group);
assertEquals(3, group.scoreDocs.length);
assertEquals(0, group.scoreDocs[0].doc);
assertEquals(1, group.scoreDocs[1].doc);
assertEquals(2, group.scoreDocs[2].doc);
assertTrue(group.scoreDocs[0].score > group.scoreDocs[1].score);
assertTrue(group.scoreDocs[1].score > group.scoreDocs[2].score);
group = groups.groups[2];
compareGroupValue("author2", group);
assertEquals(1, group.scoreDocs.length);
assertEquals(3, group.scoreDocs[0].doc);
group = groups.groups[3];
compareGroupValue(null, group);
assertEquals(1, group.scoreDocs.length);
assertEquals(6, group.scoreDocs[0].doc);
Filter lastDocInBlock = new QueryWrapperFilter(new TermQuery(new Term("groupend", "x")));
groupingSearch = new GroupingSearch(lastDocInBlock);
groups =
groupingSearch.search(indexSearcher, new TermQuery(new Term("content", "random")), 0, 10);
assertEquals(7, groups.totalHitCount);
assertEquals(7, groups.totalGroupedHitCount);
assertEquals(4, groups.totalGroupCount.longValue());
assertEquals(4, groups.groups.length);
indexSearcher.getIndexReader().close();
dir.close();
}
public void testRandomDiscreteMultiValueHighlighting() throws Exception {
String[] randomValues = new String[3 + random().nextInt(10 * RANDOM_MULTIPLIER)];
for (int i = 0; i < randomValues.length; i++) {
String randomValue;
do {
randomValue = _TestUtil.randomSimpleString(random());
} while ("".equals(randomValue));
randomValues[i] = randomValue;
}
Directory dir = newDirectory();
RandomIndexWriter writer =
new RandomIndexWriter(
random(),
dir,
newIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(random()))
.setMergePolicy(newLogMergePolicy()));
FieldType customType = new FieldType(TextField.TYPE_STORED);
customType.setStoreTermVectors(true);
customType.setStoreTermVectorOffsets(true);
customType.setStoreTermVectorPositions(true);
int numDocs = randomValues.length * 5;
int numFields = 2 + random().nextInt(5);
int numTerms = 2 + random().nextInt(3);
List<Doc> docs = new ArrayList<Doc>(numDocs);
List<Document> documents = new ArrayList<Document>(numDocs);
Map<String, Set<Integer>> valueToDocId = new HashMap<String, Set<Integer>>();
for (int i = 0; i < numDocs; i++) {
Document document = new Document();
String[][] fields = new String[numFields][numTerms];
for (int j = 0; j < numFields; j++) {
String[] fieldValues = new String[numTerms];
fieldValues[0] = getRandomValue(randomValues, valueToDocId, i);
StringBuilder builder = new StringBuilder(fieldValues[0]);
for (int k = 1; k < numTerms; k++) {
fieldValues[k] = getRandomValue(randomValues, valueToDocId, i);
builder.append(' ').append(fieldValues[k]);
}
document.add(new Field(F, builder.toString(), customType));
fields[j] = fieldValues;
}
docs.add(new Doc(fields));
documents.add(document);
}
writer.addDocuments(documents);
writer.close();
IndexReader reader = DirectoryReader.open(dir);
try {
int highlightIters = 1 + random().nextInt(120 * RANDOM_MULTIPLIER);
for (int highlightIter = 0; highlightIter < highlightIters; highlightIter++) {
String queryTerm = randomValues[random().nextInt(randomValues.length)];
int randomHit = valueToDocId.get(queryTerm).iterator().next();
List<StringBuilder> builders = new ArrayList<StringBuilder>();
for (String[] fieldValues : docs.get(randomHit).fieldValues) {
StringBuilder builder = new StringBuilder();
boolean hit = false;
for (int i = 0; i < fieldValues.length; i++) {
if (queryTerm.equals(fieldValues[i])) {
builder.append("<b>").append(queryTerm).append("</b>");
hit = true;
} else {
builder.append(fieldValues[i]);
}
if (i != fieldValues.length - 1) {
builder.append(' ');
}
}
if (hit) {
builders.add(builder);
}
}
FieldQuery fq = new FieldQuery(tq(queryTerm), true, true);
FieldTermStack stack = new FieldTermStack(reader, randomHit, F, fq);
FieldPhraseList fpl = new FieldPhraseList(stack, fq);
SimpleFragListBuilder sflb = new SimpleFragListBuilder(100);
FieldFragList ffl = sflb.createFieldFragList(fpl, 300);
SimpleFragmentsBuilder sfb = new SimpleFragmentsBuilder();
sfb.setDiscreteMultiValueHighlighting(true);
String[] actualFragments = sfb.createFragments(reader, randomHit, F, ffl, numFields);
assertEquals(builders.size(), actualFragments.length);
for (int i = 0; i < actualFragments.length; i++) {
assertEquals(builders.get(i).toString(), actualFragments[i]);
}
}
} finally {
reader.close();
dir.close();
}
}
