[SPARK-10767][PYSPARK] Make pyspark shared params codegen more consistent

[holdenk]

Namely "." shows up in some places in the template when using the param docstring and not in others

[SparkQA]

Test build #43371 has finished for PR 9017 at commit 053f3e6.

• This patch fails PySpark unit tests.
• This patch merges cleanly.
• This patch adds the following public classes (experimental):
  • class HasWeightCol(Params):
  • class IsotonicRegression(JavaEstimator, HasFeaturesCol, HasLabelCol, HasPredictionCol,
  • class IsotonicRegressionModel(JavaModel):
  • case class Average(child: Expression) extends DeclarativeAggregate
  • case class Count(child: Expression) extends DeclarativeAggregate
  • case class First(child: Expression) extends DeclarativeAggregate
  • case class Last(child: Expression) extends DeclarativeAggregate
  • case class Max(child: Expression) extends DeclarativeAggregate
  • case class Min(child: Expression) extends DeclarativeAggregate
  • abstract class StddevAgg(child: Expression) extends DeclarativeAggregate
  • case class Sum(child: Expression) extends DeclarativeAggregate
  • sealed abstract class AggregateFunction2 extends Expression with ImplicitCastInputTypes
  • abstract class ImperativeAggregate extends AggregateFunction2
  • case class RoundRobinPartitioning(numPartitions: Int) extends Partitioning
  • case class Coalesce(numPartitions: Int, child: SparkPlan) extends UnaryNode

[mengxr]

Shall we add . to the original doc instead of here?

[holdenk]

I did it the other way since it would be less changes, but yah that sounds like this would be more readable. I'll swap it :)

[SparkQA]

Test build #43409 has finished for PR 9017 at commit 1c55db7.

• This patch fails PySpark unit tests.
• This patch merges cleanly.
• This patch adds the following public classes (experimental):
  • class HasWeightCol(Params):
  • class IsotonicRegression(JavaEstimator, HasFeaturesCol, HasLabelCol, HasPredictionCol,
  • class IsotonicRegressionModel(JavaModel):
  • case class Average(child: Expression) extends DeclarativeAggregate
  • case class Count(child: Expression) extends DeclarativeAggregate
  • case class First(child: Expression) extends DeclarativeAggregate
  • case class Last(child: Expression) extends DeclarativeAggregate
  • case class Max(child: Expression) extends DeclarativeAggregate
  • case class Min(child: Expression) extends DeclarativeAggregate
  • abstract class StddevAgg(child: Expression) extends DeclarativeAggregate
  • case class Sum(child: Expression) extends DeclarativeAggregate
  • sealed abstract class AggregateFunction2 extends Expression with ImplicitCastInputTypes
  • abstract class ImperativeAggregate extends AggregateFunction2
  • case class RoundRobinPartitioning(numPartitions: Int) extends Partitioning
  • case class Coalesce(numPartitions: Int, child: SparkPlan) extends UnaryNode

[SparkQA]

Test build #43411 has finished for PR 9017 at commit 6fef2e4.

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

[SparkQA]

Test build #43447 has finished for PR 9017 at commit 929d850.

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

[holdenk]

ping @mengxr since I've applied the requested change.

[holdenk]

ping @davies since pyspark & @jkbradley since +1ed interest in having this fixed

[mengxr]

LGTM. Merged into master. Sorry for my slow response!