Tskit: Add more tests for genetic_relatedness

[brieuclehmann]

From @petrelharp on slack:

• write some more tests (e.g., for the branch and node methods), and incorporate these into the testing framework for the other statistics.

[jeromekelleher]

Comment originall in #973:

We also want to merge the current tests for genetic_relatedness into the existing framework for stats tests. Hopefully this can be done by following the pattern for something like trait regression

I'd imagine that we'd basically be moving everything out of the current test_covariance.py module, but maybe not.

[brieuclehmann]

From #975 to add test in c/tests/trees.c line 1351

• check the value of result here also. Can you put in an assert CU_ASSERT_DOUBLE_EQUAL here with the required value (fine to printf this out first to get the value - we just want to catch problems early if there's ever any regressions).

[brieuclehmann]

Also add tests for different cases of proportion, particularly the case when proportion=True. (What to do when there are no segregating sites...?)

[benjeffery]

@brieuclehmann We're considering releasing tskit Python 0.3.3 next week - do you think we can get this in? No worries if not as we can un-document this API for 0.3.3 and add that back in for 0.3.4. Thanks!

[brieuclehmann]

@benjeffery Will this be early or late next week? Later in the week may be doable but early in the week may be too much of an ask...

[benjeffery]

End of week is fine. No pressure though as I don't expect 0.3.4 will be more than a month or two away.

[brieuclehmann]

Ah right, in that case I think aiming for 0.3.4 would be safer (next week is already looking rather busy for me 😬 )

[benjeffery]

0.3.3 is held on #987 so there's more time to keep this in if you like. 0.3.4 is fine too though.

[brieuclehmann]

Thanks, I've scheduled some time to look at this in the afternoon so I should have a better idea of how long it will take later today.