rahil-c commented on code in PR #18432: URL: https://github.com/apache/hudi/pull/18432#discussion_r3024750025
########## hudi-spark-datasource/hudi-spark-common/src/main/scala/org/apache/spark/sql/hudi/analysis/VectorDistanceUtils.scala: ########## @@ -0,0 +1,159 @@ +/* + * 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.spark.sql.hudi.analysis + +import org.apache.spark.sql.catalyst.plans.logical.HoodieVectorSearchTableValuedFunction.DistanceMetric +import org.apache.spark.sql.expressions.UserDefinedFunction +import org.apache.spark.sql.functions.udf +import org.apache.spark.sql.hudi.command.exception.HoodieAnalysisException +import org.apache.spark.sql.types.{ByteType, DataType, DoubleType, FloatType} + +/** + * Vector distance utilities: raw distance functions and Spark UDF factories. + * + * Raw functions operate on Array[Double] and can be used outside of Spark + * (e.g. for verification, index building, or non-Spark distance computation). + * + * UDF factories produce typed Spark UDFs for Float, Double, and Byte corpus columns. + */ +object VectorDistanceUtils { Review Comment: Let me look into this more in regards to some existing libraries for just doing the distance calculation for cosine, l2, dot prod etc. When it comes to high performance libraries that can do this vector search this PR just adds a basic brute force, however we plan as a fast follow to ideally implement two other approaches from spark libs such a `RowMatrix` approach and leveraging `spark-hnsw`. Finally I think the scope of RFC 102 was a spark TVF so we I think for now for 1.2 were scoping this to just be focused on spark as the primary engine. -- 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: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
