Adds Iterable<DataRowSchema>.toDataFrame() overload

[Jolanrensen]

Fixes #1880

Added Iterable.toDataFrame() overload for DataRowSchema types,
compiler plugin test in :samples, because the compiler plugin is live there.

This fixes the compiler plugin creating a new type immediately when turning a list of @DataSchema annotated classes into a dataframe.

[Jolanrensen]

As expected, this directly touches the kotlin-spark dev example :D (where this limitation was detected) I'll fix the example right away

[Jolanrensen]

A workaround for #1908 could be to write the function like this:

@JvmName("dataRowSchemaIterableToDataFrame")
@Suppress("INCORRECT_LEFT_COMPONENT_OF_INTERSECTION")
public inline fun <reified T : DataRowSchema> Iterable<T>.toDataFrame(): DataFrame<T & Any> =
    @DisableInterpretation
    toDataFrame {
        properties()
    }

But notebooks don't like T & Any and it could better be solved from the compiler plugin side I believe.

[koperagen]

Let's probably not merge since this needs a workaround if flexible type appears. A few things need to be fixed on compiler plugin side to properly work with such types. Let's wait these fixes in stable release of Kotlin, postpone feature to 1.1

Even this is not ideal because it suppresses INCORRECT_LEFT_COMPONENT_OF_INTERSECTION, which is not guaranteed to work in the future

public inline fun <reified T : DataRowSchema> Iterable<T>.toDataFrame(): DataFrame<T & Any> =
    @DisableInterpretation
    toDataFrame {
        properties()
    }