Resampling for univariate KS test

At the moment the KS testing package implemented in the `similarity` module is easy to pass if the sample of observational data is small (which it often is). We might want to see whether resampling gives different results than the current implementation of the univariate KS testing in `xks`.

Start by reading the `similarity` module and the current univariate KS testing implementation in `xks`. Investigate how resampling could be compared with the existing behavior for small observational samples; the work is done when the effect of resampling is evaluated and the resulting approach is clearly documented.

Written by the indexing model from the issue text.