* Minor bug-fix: No need to raise error when using alternative correction methods, since we return None CIs and print out warning instead
2.3.1
------------------
* Added even more multiple correction strategies
2.3.0
------------------
* Added additional multiple correction methods (Holm, Hommel, Simes-Hochberg) for one sided tests. * Added verbose mode to summary and difference methods that returns all intermediate columns that are used in the computations
2.2.0
------------------
* Changed how sequential tests are done. Now, instead of passing a single (number) final_expected_sample_size, you pass a column name final_expected_sample_size. This is to ensure that groupby works as expected, i.e. different groups can have different expected sample sizes.
2.1.4
------------------
* Added support for NaNs in NIMs
2.1.3
------------------
* Added option to pass non_inferiority_margins=True which then uses NIMs in source data_frame, rather than passing in dict of tuples with NIMs