ENH: Support kurtosis (kurt) in DataFrameGroupBy and SeriesGroupBy

[snitish]

• closes ENH:AttributeError: 'SeriesGroupBy' object has no attribute 'kurtosis' #40139
• Tests added and passed if fixing a bug or adding a new feature
• All code checks passed.
• Added type annotations to new arguments/methods/functions.
• Added an entry in the latest doc/source/whatsnew/vX.X.X.rst file if fixing a bug or adding a new feature.

DataFrameGroupBy and SeriesGroupBy currently support mean, std and skew (the first 3 moments) but not kurtosis (the 4th moment). This change addresses that. I implemented kurtosis in cython in similar fashion to skewness. I've verified that the output of this function matches that of DataFrame.kurt().

[rhshadrach]

Very nice! A few test requests (I think these are not covered yet):

• skipna with NA values in the data
• Float64 and float64[pyarrow] dtypes
• Constant data (e.g. [1, 1, 1, 1])

[snitish]

Thanks for the review @rhshadrach.

• Addressed your comments
• Added test case for skipna=False (by default it's true)
• Added test case for float64[pyarrow] (we already have one for float64)
• Added test case for constant data. Note that the result here is 0.0, consistent with DataFrame.kurt() and Series.kurt()