[SPARK-29258][ML][PYSPARK] parity between ml.evaluator and mllib.metrics

[zhengruifeng]

What changes were proposed in this pull request?

1, expose BinaryClassificationMetrics.numBins in BinaryClassificationEvaluator
2, expose RegressionMetrics.throughOrigin in RegressionEvaluator
3, add metric explainedVariance in RegressionEvaluator

Why are the changes needed?

existing function in mllib.metrics should also be exposed in ml

Does this PR introduce any user-facing change?

yes, this PR add two expert params and one metric option

How was this patch tested?

existing and added tests

[SparkQA]

Test build #111411 has finished for PR 25940 at commit aaef3f7.

• This patch fails PySpark unit tests.
• This patch merges cleanly.
• This patch adds no public classes.

[SparkQA]

Test build #111422 has finished for PR 25940 at commit 932a17e.

• This patch fails PySpark unit tests.
• This patch merges cleanly.
• This patch adds no public classes.

[SparkQA]

Test build #111424 has finished for PR 25940 at commit 4b4cbc6.

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

[srowen]

Total nit, but you could write case "r2" | "var" => true

[SparkQA]

Test build #111459 has finished for PR 25940 at commit 45139b5.

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

[zhengruifeng]

Merged to master. Thanks @srowen for reviewing