Provide 2-sample Anderson Darling test as constraint
The recently implemented Kolmogorov Smirnov test can be useful when trying to assess whether two samples are drawn from the identical underlying distribution. While it might already satisfy a user's need, some might argue (e.g. [0]) that other kinds of tests might be more adequate in certain circumstances. A prominent alternative could be the Anderson Darling [1] test.
Once our Kolmogorov Smirnov test query logic is implemented via the sqlalchemy language expression api - see https://github.com/Quantco/datajudge/issues/29 - it should be very easy to share a lot of the underlying query logic between both tests.
[0] https://www.researchgate.net/publication/267205556_Power_Comparisons_of_Shapiro-Wilk_Kolmogorov-Smirnov_Lilliefors_and_Anderson-Darling_Tests [1] https://en.wikipedia.org/wiki/Anderson%E2%80%93Darling_test#Non-parametric_k-sample_tests