uschindler commented on code in PR #12253: URL: https://github.com/apache/lucene/pull/12253#discussion_r1226849047
########## lucene/queries/src/java/org/apache/lucene/queries/function/valuesource/KnnVectorFieldSource.java: ########## @@ -0,0 +1,112 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ +package org.apache.lucene.queries.function.valuesource; + +import java.io.IOException; +import java.util.Map; +import org.apache.lucene.index.ByteVectorValues; +import org.apache.lucene.index.FloatVectorValues; +import org.apache.lucene.index.LeafReaderContext; +import org.apache.lucene.queries.function.FunctionValues; +import org.apache.lucene.queries.function.ValueSource; +import org.apache.lucene.search.DocIdSetIterator; + +/** An implementation for retrieving {@link FunctionValues} instances for knn vectors fields. */ +public class KnnVectorFieldSource extends ValueSource { + private final String fieldName; + + public KnnVectorFieldSource(String fieldName) { + this.fieldName = fieldName; + } + + @Override + public FunctionValues getValues(Map<Object, Object> context, LeafReaderContext readerContext) + throws IOException { + + final FloatVectorValues floatValues = readerContext.reader().getFloatVectorValues(fieldName); + final ByteVectorValues byteValues = readerContext.reader().getByteVectorValues(fieldName); + + if (floatValues == null && byteValues == null) { + throw new IllegalArgumentException( + "no vector value is indexed for field '" + fieldName + "'"); + } + return new FunctionValues() { + int lastDocID; + + @Override + public float[] floatVectorVal(int doc) throws IOException { + float[] floatVector = null; + if (floatValues != null && exists(floatValues, doc)) { + floatVector = floatValues.vectorValue(); + } else { + byte[] byteVector = byteVectorVal(doc); + if (byteVector != null) { + floatVector = new float[byteVector.length]; + for (int i = 0; i < byteVector.length; i++) { + floatVector[i] = byteVector[i]; + } + } + } + return floatVector; + } Review Comment: I fully agree with @benwtrent. We also have separate queries for byte and float vectors, so we should not mix the types together in function queries. It wil really hurt users. I know Solr people tend to sometimes "try to change their schema", but this is a no-go for byte and float vectors as other queries will fail but this one suddenly works magically. No, we should not do this. Field types in Lucene were relaxed at beginning but we spent more and more time into making them more strict. We only have some autoconverts in docvalues (e.g., SortedDocValues can be queries and sorted using SortedSetDocValues which simplifies code - bt there is no surprises and loss or convert of information). -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org For additional commands, e-mail: issues-h...@lucene.apache.org
