Add kmeans clustering function

[oyvindeide]

Closes #241

[eivindjahren]

I think in this case, you get a cleaner implementation with hypothesis than with pytest parametrization.
Just a suggestion:

from hypothesis import given, assume
import hypothesis.strategies as st

clustering_functions = st.sampled_from(["hierarchical", "kmeans"])
methods = st.sampled_from(["spearman_correlation", "auto_scale"])

@given(st.integers(min_value=1, max_value=100), methods, clustering_functions)
def test_misfit_preprocessor_n_polynomials(
   num_polynomials, method, clustering_function
):
    if (
        clustering_function == "kmeans"
        and method == "spearman_correlation"
     ):
       assume(num_polynomials in [4,5]) # or more

[lars-petter-hauge]

Had some minor comments, otherwise I think the implementation looks good!

[lars-petter-hauge]

LGTM! Nice job!