[SPARK-11029] [ML] Add computeCost to KMeansModel in spark.ml

[hhbyyh]

jira: https://issues.apache.org/jira/browse/SPARK-11029

We should add a method analogous to spark.mllib.clustering.KMeansModel.computeCost to spark.ml.clustering.KMeansModel.
This will be a temp fix until we have proper evaluators defined for clustering.

[SparkQA]

Test build #43562 has finished for PR 9073 at commit 5440dc5.

• This patch passes all tests.
• This patch merges cleanly.
• This patch adds no public classes.

[jkbradley]

@hhbyyh I think this should take a DataFrame instead of an RDD to be consistent with the Pipelines API. Also, would it be easy to add a quick unit test? Thanks!

[hhbyyh]

@jkbradley Thanks for helping review. Let me know if we need a separate UT.

[SparkQA]

Test build #43593 has finished for PR 9073 at commit 738ab03.

• This patch passes all tests.
• This patch merges cleanly.
• This patch adds no public classes.

[jkbradley]

no longer necessary

[jkbradley]

That should be it---thanks!

[jkbradley]

LGTM pending tests

[SparkQA]

Test build #43864 has finished for PR 9073 at commit cff3d5a.

• This patch passes all tests.
• This patch merges cleanly.
• This patch adds no public classes.

[hhbyyh]

Thanks for taking the time review.

[jkbradley]

Merging with master. Thanks for the PR!