Github user srowen commented on a diff in the pull request:
https://github.com/apache/spark/pull/20600#discussion_r171625847
--- Diff: project/MimaExcludes.scala ---
@@ -36,6 +36,12 @@ object MimaExcludes {
// Exclude rules for 2.4.x
lazy val v24excludes = v23excludes ++ Seq(
+ // [SPARK-23412][ML] Add cosine distance measure to BisectingKmeans
+
ProblemFilters.exclude[InheritedNewAbstractMethodProblem]("org.apache.spark.ml.param.shared.HasDistanceMeasure.org$apache$spark$ml$param$shared$HasDistanceMeasure$_setter_$distanceMeasure_="),
+
ProblemFilters.exclude[InheritedNewAbstractMethodProblem]("org.apache.spark.ml.param.shared.HasDistanceMeasure.getDistanceMeasure"),
+
ProblemFilters.exclude[InheritedNewAbstractMethodProblem]("org.apache.spark.ml.param.shared.HasDistanceMeasure.distanceMeasure"),
+
ProblemFilters.exclude[DirectMissingMethodProblem]("org.apache.spark.mllib.clustering.BisectingKMeansModel#SaveLoadV1_0.load"),
--- End diff --
See above; this one I think needs to be avoided. The others are issues too
but probably acceptable in a minor release. I don't know if people extend these
implementations. That said, any way to work around this by providing a dummy
implementation of the new methods? I haven't thought through the Scala here.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
