Skip to content

[SPARK-25793][ML]call SaveLoadV2_0.load for classNameV2_0 #22790

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
huaxingao wants to merge 4 commits into from
Closed

[SPARK-25793][ML]call SaveLoadV2_0.load for classNameV2_0 #22790

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
d599f83
[SPARK-25793][ML]call SaveLoadV2_0.load for classNameV2_0
huaxingao Oct 22, 2018
77902bc
change BisectingKMeansModel.save to use SaveLoadV2_0.save
huaxingao Oct 23, 2018
29c9ca0
add test
huaxingao Oct 23, 2018
0decb23
change formatVersion to 2.0
huaxingao Oct 25, 2018
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 3 additions & 3 deletions mllib/src/main/scala/org/apache/spark/mllib/clustering/BisectingKMeansModel.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -109,10 +109,10 @@ class BisectingKMeansModel private[clustering] (

@Since("2.0.0")
override def save(sc: SparkContext, path: String): Unit = {
BisectingKMeansModel.SaveLoadV1_0.save(sc, this, path)
BisectingKMeansModel.SaveLoadV2_0.save(sc, this, path)
}

override protected def formatVersion: String = "1.0"
override protected def formatVersion: String = "2.0"
}

@Since("2.0.0")
Expand All @@ -126,7 +126,7 @@ object BisectingKMeansModel extends Loader[BisectingKMeansModel] {
val model = SaveLoadV1_0.load(sc, path)
model
case (SaveLoadV2_0.thisClassName, SaveLoadV2_0.thisFormatVersion) =>
val model = SaveLoadV1_0.load(sc, path)
val model = SaveLoadV2_0.load(sc, path)
Copy link
Member

@viirya viirya Oct 22, 2018

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

cc @mgaido91

Copy link
Contributor

@mgaido91 mgaido91 Oct 22, 2018

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

ah, nice catch!

Copy link
Member

@dongjoon-hyun dongjoon-hyun Oct 22, 2018 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This is not a regression, but it looks like a correctness or data loss issue at Spark 2.4.0 new feature.
Do you know why https://issues.apache.org/jira/browse/SPARK-25793 is created as a Minor and Documentation issue?

Copy link
Member

@viirya viirya Oct 23, 2018

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

cc @cloud-fan

Copy link
Member

@viirya viirya Oct 22, 2018

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Is there no test to verify calling correct load method?

Copy link
Contributor

@mgaido91 mgaido91 Oct 22, 2018

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think we can improve the write/load model tests in order to include also a different distance measure from the default one. In this way we should catch this error. Thanks.

Copy link
Member

@viirya viirya Oct 23, 2018

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Do we have ever use SaveLoadV2_0 to save model for now? Looks BisectingKMeansModel.save simply calls SaveLoadV1_0.save to save models.

Copy link
Contributor Author

@huaxingao huaxingao Oct 23, 2018

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@viirya @mgaido91
I will change the BisectingKMeansModel.save to

BisectingKMeansModel.SaveLoadV2_0.save(sc, this, path)

model
case _ => throw new Exception(
s"BisectingKMeansModel.load did not recognize model with (className, format version):" +
Expand Down
3 changes: 2 additions & 1 deletion mllib/src/test/scala/org/apache/spark/mllib/clustering/BisectingKMeansSuite.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -187,11 +187,12 @@ class BisectingKMeansSuite extends SparkFunSuite with MLlibTestSparkContext {

val points = (1 until 8).map(i => Vectors.dense(i))
val data = sc.parallelize(points, 2)
val model = new BisectingKMeans().run(data)
val model = new BisectingKMeans().setDistanceMeasure(DistanceMeasure.COSINE).run(data)
try {
model.save(sc, path)
val sameModel = BisectingKMeansModel.load(sc, path)
assert(model.k === sameModel.k)
assert(model.distanceMeasure === sameModel.distanceMeasure)
model.clusterCenters.zip(sameModel.clusterCenters).foreach(c => c._1 === c._2)
} finally {
Utils.deleteRecursively(tempDir)
Expand Down