private static String resolveIndexName(String fieldName, QueryShardContext context) {
MappedFieldType fieldType = context.fieldMapper(fieldName);
if (fieldType != null) {
return fieldType.names().indexName();
}
return fieldName;
}
@Override
public IndexFieldData<AtomicNumericFieldData> build(
Index index,
@IndexSettings Settings indexSettings,
MappedFieldType fieldType,
IndexFieldDataCache cache,
CircuitBreakerService breakerService,
MapperService mapperService) {
return new PackedArrayIndexFieldData(
index,
indexSettings,
fieldType.names(),
fieldType.fieldDataType(),
cache,
numericType,
breakerService);
}
@Override
protected Query doToQuery(QueryShardContext context) throws IOException {
String field;
MappedFieldType fieldType = context.fieldMapper(fieldName);
if (fieldType != null) {
field = fieldType.names().indexName();
} else {
field = fieldName;
}
Analyzer analyzerObj;
if (analyzer == null) {
if (fieldType != null) {
analyzerObj = context.getSearchAnalyzer(fieldType);
} else {
analyzerObj = context.getMapperService().searchAnalyzer();
}
} else {
analyzerObj = context.getMapperService().analysisService().analyzer(analyzer);
if (analyzerObj == null) {
throw new QueryShardException(context, "[common] analyzer [" + analyzer + "] not found");
}
}
Occur highFreqOccur = highFreqOperator.toBooleanClauseOccur();
Occur lowFreqOccur = lowFreqOperator.toBooleanClauseOccur();
ExtendedCommonTermsQuery commonsQuery =
new ExtendedCommonTermsQuery(
highFreqOccur, lowFreqOccur, cutoffFrequency, disableCoord, fieldType);
return parseQueryString(
commonsQuery,
text,
field,
analyzerObj,
lowFreqMinimumShouldMatch,
highFreqMinimumShouldMatch);
}
@Override
public void execute(SearchContext context) {
FieldsVisitor fieldsVisitor;
Set<String> fieldNames = null;
List<String> extractFieldNames = null;
boolean loadAllStored = false;
if (!context.hasFieldNames()) {
// no fields specified, default to return source if no explicit indication
if (!context.hasScriptFields() && !context.hasFetchSourceContext()) {
context.fetchSourceContext(new FetchSourceContext(true));
}
fieldsVisitor = new FieldsVisitor(context.sourceRequested());
} else if (context.fieldNames().isEmpty()) {
fieldsVisitor = new FieldsVisitor(context.sourceRequested());
} else {
for (String fieldName : context.fieldNames()) {
if (fieldName.equals("*")) {
loadAllStored = true;
continue;
}
if (fieldName.equals(SourceFieldMapper.NAME)) {
if (context.hasFetchSourceContext()) {
context.fetchSourceContext().fetchSource(true);
} else {
context.fetchSourceContext(new FetchSourceContext(true));
}
continue;
}
MappedFieldType fieldType = context.smartNameFieldType(fieldName);
if (fieldType == null) {
// Only fail if we know it is a object field, missing paths / fields shouldn't fail.
if (context.getObjectMapper(fieldName) != null) {
throw new IllegalArgumentException("field [" + fieldName + "] isn't a leaf field");
}
} else if (fieldType.stored()) {
if (fieldNames == null) {
fieldNames = new HashSet<>();
}
fieldNames.add(fieldType.names().indexName());
} else {
if (extractFieldNames == null) {
extractFieldNames = newArrayList();
}
extractFieldNames.add(fieldName);
}
}
if (loadAllStored) {
fieldsVisitor = new AllFieldsVisitor(); // load everything, including _source
} else if (fieldNames != null) {
boolean loadSource = extractFieldNames != null || context.sourceRequested();
fieldsVisitor = new CustomFieldsVisitor(fieldNames, loadSource);
} else {
fieldsVisitor = new FieldsVisitor(extractFieldNames != null || context.sourceRequested());
}
}
InternalSearchHit[] hits = new InternalSearchHit[context.docIdsToLoadSize()];
FetchSubPhase.HitContext hitContext = new FetchSubPhase.HitContext();
for (int index = 0; index < context.docIdsToLoadSize(); index++) {
int docId = context.docIdsToLoad()[context.docIdsToLoadFrom() + index];
int readerIndex = ReaderUtil.subIndex(docId, context.searcher().getIndexReader().leaves());
LeafReaderContext subReaderContext =
context.searcher().getIndexReader().leaves().get(readerIndex);
int subDocId = docId - subReaderContext.docBase;
final InternalSearchHit searchHit;
try {
int rootDocId = findRootDocumentIfNested(context, subReaderContext, subDocId);
if (rootDocId != -1) {
searchHit =
createNestedSearchHit(
context,
docId,
subDocId,
rootDocId,
extractFieldNames,
loadAllStored,
fieldNames,
subReaderContext);
} else {
searchHit =
createSearchHit(
context, fieldsVisitor, docId, subDocId, extractFieldNames, subReaderContext);
}
} catch (IOException e) {
throw ExceptionsHelper.convertToElastic(e);
}
hits[index] = searchHit;
hitContext.reset(searchHit, subReaderContext, subDocId, context.searcher());
for (FetchSubPhase fetchSubPhase : fetchSubPhases) {
if (fetchSubPhase.hitExecutionNeeded(context)) {
fetchSubPhase.hitExecute(context, hitContext);
}
}
}
for (FetchSubPhase fetchSubPhase : fetchSubPhases) {
if (fetchSubPhase.hitsExecutionNeeded(context)) {
fetchSubPhase.hitsExecute(context, hits);
}
}
context
.fetchResult()
.hits(
new InternalSearchHits(
hits,
context.queryResult().topDocs().totalHits,
context.queryResult().topDocs().getMaxScore()));
}
@Override
public Query parse(QueryParseContext parseContext) throws IOException, QueryParsingException {
XContentParser parser = parseContext.parser();
MoreLikeThisQuery mltQuery = new MoreLikeThisQuery();
mltQuery.setSimilarity(parseContext.searchSimilarity());
Analyzer analyzer = null;
List<String> moreLikeFields = null;
boolean failOnUnsupportedField = true;
String queryName = null;
boolean include = false;
XContentParser.Token token;
String currentFieldName = null;
List<String> likeTexts = new ArrayList<>();
MultiTermVectorsRequest likeItems = new MultiTermVectorsRequest();
List<String> unlikeTexts = new ArrayList<>();
MultiTermVectorsRequest unlikeItems = new MultiTermVectorsRequest();
while ((token = parser.nextToken()) != XContentParser.Token.END_OBJECT) {
if (token == XContentParser.Token.FIELD_NAME) {
currentFieldName = parser.currentName();
} else if (token.isValue()) {
if (parseContext.parseFieldMatcher().match(currentFieldName, Fields.LIKE_TEXT)) {
likeTexts.add(parser.text());
} else if (parseContext.parseFieldMatcher().match(currentFieldName, Fields.LIKE)) {
parseLikeField(parser, likeTexts, likeItems);
} else if (parseContext.parseFieldMatcher().match(currentFieldName, Fields.UNLIKE)) {
parseLikeField(parser, unlikeTexts, unlikeItems);
} else if (parseContext.parseFieldMatcher().match(currentFieldName, Fields.MIN_TERM_FREQ)) {
mltQuery.setMinTermFrequency(parser.intValue());
} else if (parseContext
.parseFieldMatcher()
.match(currentFieldName, Fields.MAX_QUERY_TERMS)) {
mltQuery.setMaxQueryTerms(parser.intValue());
} else if (parseContext.parseFieldMatcher().match(currentFieldName, Fields.MIN_DOC_FREQ)) {
mltQuery.setMinDocFreq(parser.intValue());
} else if (parseContext.parseFieldMatcher().match(currentFieldName, Fields.MAX_DOC_FREQ)) {
mltQuery.setMaxDocFreq(parser.intValue());
} else if (parseContext
.parseFieldMatcher()
.match(currentFieldName, Fields.MIN_WORD_LENGTH)) {
mltQuery.setMinWordLen(parser.intValue());
} else if (parseContext
.parseFieldMatcher()
.match(currentFieldName, Fields.MAX_WORD_LENGTH)) {
mltQuery.setMaxWordLen(parser.intValue());
} else if (parseContext.parseFieldMatcher().match(currentFieldName, Fields.BOOST_TERMS)) {
float boostFactor = parser.floatValue();
if (boostFactor != 0) {
mltQuery.setBoostTerms(true);
mltQuery.setBoostTermsFactor(boostFactor);
}
} else if (parseContext
.parseFieldMatcher()
.match(currentFieldName, Fields.MINIMUM_SHOULD_MATCH)) {
mltQuery.setMinimumShouldMatch(parser.text());
} else if ("analyzer".equals(currentFieldName)) {
analyzer = parseContext.analysisService().analyzer(parser.text());
} else if ("boost".equals(currentFieldName)) {
mltQuery.setBoost(parser.floatValue());
} else if (parseContext
.parseFieldMatcher()
.match(currentFieldName, Fields.FAIL_ON_UNSUPPORTED_FIELD)) {
failOnUnsupportedField = parser.booleanValue();
} else if ("_name".equals(currentFieldName)) {
queryName = parser.text();
} else if (parseContext.parseFieldMatcher().match(currentFieldName, Fields.INCLUDE)) {
include = parser.booleanValue();
} else {
throw new QueryParsingException(
parseContext, "[mlt] query does not support [" + currentFieldName + "]");
}
} else if (token == XContentParser.Token.START_ARRAY) {
if (parseContext.parseFieldMatcher().match(currentFieldName, Fields.STOP_WORDS)) {
Set<String> stopWords = Sets.newHashSet();
while ((token = parser.nextToken()) != XContentParser.Token.END_ARRAY) {
stopWords.add(parser.text());
}
mltQuery.setStopWords(stopWords);
} else if ("fields".equals(currentFieldName)) {
moreLikeFields = new LinkedList<>();
while ((token = parser.nextToken()) != XContentParser.Token.END_ARRAY) {
String field = parser.text();
MappedFieldType fieldType = parseContext.fieldMapper(field);
moreLikeFields.add(fieldType == null ? field : fieldType.names().indexName());
}
} else if (parseContext.parseFieldMatcher().match(currentFieldName, Fields.DOCUMENT_IDS)) {
while ((token = parser.nextToken()) != XContentParser.Token.END_ARRAY) {
if (!token.isValue()) {
throw new IllegalArgumentException("ids array element should only contain ids");
}
likeItems.add(newTermVectorsRequest().id(parser.text()));
}
} else if (parseContext.parseFieldMatcher().match(currentFieldName, Fields.DOCUMENTS)) {
while ((token = parser.nextToken()) != XContentParser.Token.END_ARRAY) {
if (token != XContentParser.Token.START_OBJECT) {
throw new IllegalArgumentException("docs array element should include an object");
}
likeItems.add(parseDocument(parser));
}
} else if (parseContext.parseFieldMatcher().match(currentFieldName, Fields.LIKE)) {
while ((token = parser.nextToken()) != XContentParser.Token.END_ARRAY) {
parseLikeField(parser, likeTexts, likeItems);
}
} else if (parseContext.parseFieldMatcher().match(currentFieldName, Fields.UNLIKE)) {
while ((token = parser.nextToken()) != XContentParser.Token.END_ARRAY) {
parseLikeField(parser, unlikeTexts, unlikeItems);
}
} else {
throw new QueryParsingException(
parseContext, "[mlt] query does not support [" + currentFieldName + "]");
}
} else if (token == XContentParser.Token.START_OBJECT) {
if (parseContext.parseFieldMatcher().match(currentFieldName, Fields.LIKE)) {
parseLikeField(parser, likeTexts, likeItems);
} else if (parseContext.parseFieldMatcher().match(currentFieldName, Fields.UNLIKE)) {
parseLikeField(parser, unlikeTexts, unlikeItems);
} else {
throw new QueryParsingException(
parseContext, "[mlt] query does not support [" + currentFieldName + "]");
}
}
}
if (likeTexts.isEmpty() && likeItems.isEmpty()) {
throw new QueryParsingException(
parseContext, "more_like_this requires 'like' to be specified");
}
if (moreLikeFields != null && moreLikeFields.isEmpty()) {
throw new QueryParsingException(
parseContext, "more_like_this requires 'fields' to be non-empty");
}
// set analyzer
if (analyzer == null) {
analyzer = parseContext.mapperService().searchAnalyzer();
}
mltQuery.setAnalyzer(analyzer);
// set like text fields
boolean useDefaultField = (moreLikeFields == null);
if (useDefaultField) {
moreLikeFields = Collections.singletonList(parseContext.defaultField());
}
// possibly remove unsupported fields
removeUnsupportedFields(moreLikeFields, analyzer, failOnUnsupportedField);
if (moreLikeFields.isEmpty()) {
return null;
}
mltQuery.setMoreLikeFields(moreLikeFields.toArray(Strings.EMPTY_ARRAY));
// support for named query
if (queryName != null) {
parseContext.addNamedQuery(queryName, mltQuery);
}
// handle like texts
if (!likeTexts.isEmpty()) {
mltQuery.setLikeText(likeTexts);
}
if (!unlikeTexts.isEmpty()) {
mltQuery.setIgnoreText(unlikeTexts);
}
// handle items
if (!likeItems.isEmpty()) {
// set default index, type and fields if not specified
MultiTermVectorsRequest items = likeItems;
for (TermVectorsRequest item : unlikeItems) {
items.add(item);
}
for (TermVectorsRequest item : items) {
if (item.index() == null) {
item.index(parseContext.index().name());
}
if (item.type() == null) {
if (parseContext.queryTypes().size() > 1) {
throw new QueryParsingException(
parseContext,
"ambiguous type for item with id: " + item.id() + " and index: " + item.index());
} else {
item.type(parseContext.queryTypes().iterator().next());
}
}
// default fields if not present but don't override for artificial docs
if (item.selectedFields() == null && item.doc() == null) {
if (useDefaultField) {
item.selectedFields("*");
} else {
item.selectedFields(moreLikeFields.toArray(new String[moreLikeFields.size()]));
}
}
}
// fetching the items with multi-termvectors API
items.copyContextAndHeadersFrom(SearchContext.current());
MultiTermVectorsResponse responses = fetchService.fetchResponse(items);
// getting the Fields for liked items
mltQuery.setLikeText(MoreLikeThisFetchService.getFields(responses, likeItems));
// getting the Fields for ignored items
if (!unlikeItems.isEmpty()) {
org.apache.lucene.index.Fields[] ignoreFields =
MoreLikeThisFetchService.getFields(responses, unlikeItems);
if (ignoreFields.length > 0) {
mltQuery.setUnlikeText(ignoreFields);
}
}
BooleanQuery.Builder boolQuery = new BooleanQuery.Builder();
boolQuery.add(mltQuery, BooleanClause.Occur.SHOULD);
// exclude the items from the search
if (!include) {
handleExclude(boolQuery, likeItems);
}
return boolQuery.build();
}
return mltQuery;
}
@SuppressWarnings("unchecked")
public <IFD extends IndexFieldData<?>> IFD getForField(MappedFieldType fieldType) {
final Names fieldNames = fieldType.names();
final FieldDataType type = fieldType.fieldDataType();
if (type == null) {
throw new IllegalArgumentException(
"found no fielddata type for field [" + fieldNames.fullName() + "]");
}
final boolean docValues = fieldType.hasDocValues();
IndexFieldData.Builder builder = null;
String format = type.getFormat(indexSettings);
if (format != null && FieldDataType.DOC_VALUES_FORMAT_VALUE.equals(format) && !docValues) {
logger.warn(
"field ["
+ fieldNames.fullName()
+ "] has no doc values, will use default field data format");
format = null;
}
if (format != null) {
builder = buildersByTypeAndFormat.get(Tuple.tuple(type.getType(), format));
if (builder == null) {
logger.warn(
"failed to find format ["
+ format
+ "] for field ["
+ fieldNames.fullName()
+ "], will use default");
}
}
if (builder == null && docValues) {
builder = docValuesBuildersByType.get(type.getType());
}
if (builder == null) {
builder = buildersByType.get(type.getType());
}
if (builder == null) {
throw new IllegalArgumentException(
"failed to find field data builder for field "
+ fieldNames.fullName()
+ ", and type "
+ type.getType());
}
IndexFieldDataCache cache;
synchronized (this) {
cache = fieldDataCaches.get(fieldNames.indexName());
if (cache == null) {
// we default to node level cache, which in turn defaults to be unbounded
// this means changing the node level settings is simple, just set the bounds there
String cacheType =
type.getSettings()
.get("cache", indexSettings.get(FIELDDATA_CACHE_KEY, FIELDDATA_CACHE_VALUE_NODE));
if (FIELDDATA_CACHE_VALUE_NODE.equals(cacheType)) {
cache = indicesFieldDataCache.buildIndexFieldDataCache(listener, index, fieldNames, type);
} else if ("none".equals(cacheType)) {
cache = new IndexFieldDataCache.None();
} else {
throw new IllegalArgumentException(
"cache type not supported ["
+ cacheType
+ "] for field ["
+ fieldNames.fullName()
+ "]");
}
fieldDataCaches.put(fieldNames.indexName(), cache);
}
}
return (IFD)
builder.build(index, indexSettings, fieldType, cache, circuitBreakerService, mapperService);
}
