Create more precise annotations for DataFrame/Series GroupBy operations (agg, apply, transform)

[gandhis1]

All of these groupby operations take a function that can be a str or a Callable. They should ensure that:

• The Callable accepts a first parameter that is a DataFrame or Series as applicable
• The Callable returns the appropriate type - in some cases an Iterable, in other cases either a DataFrame or Series
• If a str is passed in, it is part of the set of supported functions, which are located here: https://github.com/pandas-dev/pandas/blob/main/pandas/core/groupby/base.py#L60
• If possible - and this may not be possible today - ensure that args and kwargs parameters are type checked against the passed-in Callable
• Comprehensive set of tests are needed here, to exercise all variations

I'll start with transform because that is a fairly simple API compared to the others and it is not very commonly used, so the potential impact of more narrow types there has a lower risk of an impactful regression.

[Dr-Irv]

For transform, see this line, https://github.com/pandas-dev/pandas/blob/6b0ac938279d2baec210641bfaf3719872f9a187/pandas/core/groupby/base.py#L118 so you have to combine the transform specific ones and the reduction ones.