[SPARK-16898][SQL] Adds argument type information for typed logical plan like MapElements, TypedFilter, and AppendColumn

[clockfly]

What changes were proposed in this pull request?

This PR adds argument type information for typed logical plan like MapElements, TypedFilter, and AppendColumn, so that we can use these info in customized optimizer rule.

How was this patch tested?

Existing test.

[SparkQA]

Test build #63229 has finished for PR 14494 at commit c15b46c.

• This patch passes all tests.
• This patch merges cleanly.
• This patch adds the following public classes (experimental):
  • class TypedSumDouble[IN](val f: IN => Double) extends Aggregator[IN, Double, Double]
  • class TypedSumLong[IN](val f: IN => Long) extends Aggregator[IN, Long, Long]
  • class TypedCount[IN](val f: IN => Any) extends Aggregator[IN, Long, Long]
  • class TypedAverage[IN](val f: IN => Double) extends Aggregator[IN, (Double, Long), Double]

[cloud-fan]

why expose the f?

[clockfly]

TypedFilter and MapElements also exposes the function closure? With this, we can apply the possible optimization describe in https://issues.apache.org/jira/browse/SPARK-16898?

[cloud-fan]

thanks, merging to master!