Put data fixture for all tests into separate file to reduce code duplication

[amichuda]

Right now test_CL, test_HC, test_seeds and test_weights all have a data function.

We can put the data fixture into a conftest.py file that is global so that all tests can use the data function.

I also noticed that sometimes data() also takes a G parameter. We can use pytest.mark.parameterize to allow for multiple tests at different numbers of clusters for test robustness.