GroupBy aggregate fix

[Jolanrensen]

With the debug mode enabled, one failing test of #713 was in GroupBy.aggregate { this into "" }.

I narrowed it down to a NamedValue being created in GroupByReceiverImpl with KType AggregateGroupedDsl<DfType>. Since this value is an AggregateInternalDsl<*> too, it is unpacked in the yield() function to get the underlying dataframe (by copying the NamedValue and replacing its value with its value.df), however, the KType of this NamedValue was left unchanged and thus mismatched the value.

This PR fixes that by updating that KType value too

[Jolanrensen]

@zaleslaw the last commit can be reviewed if you like. It builds on top of #801 which should be merged first

[github-actions]

Generated sources will be updated after merging this PR.
Please inspect the changes in here.

[zaleslaw]

I'm a little bit worried about this expression with 6 ? - what if something goes wrong and will be null, I could not explain myself the logic, should we handle the alternative (or common alternative)?

[Jolanrensen]

That's what the ?: is for at the end. We try to create DataFrame<Type1> from AggregateInternalDsl<Type1>. However, in theory, any KType can be supplied by the user. KTypes like GroupByReceiverImpl<Type1> , DataFrame<Type1> (which would both still work), but unfortunately also KTypes like CompletelyCustomType without arguments. So in cases where this type argument cannot be fetched, we default to DataFrame<*> with the ?:.