Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Overall, the goal was to mirror other sample set stat design patterns, with the exception that we dispatch each low level stat matrix from the ld_matrix function. We also fix some behavior uncovered in testing: * Ensure the D_prime statistic does not divide by zero. (opened #2907 to investigate if this something we actually want to do). * Look up node from sample_index_map when finding samples under mutation nodes. * Do not error out if no sites are specified, instead return an array with (0, 0) dimensions. Remove associated test, add one for early return. Exhaustive test coverage was added, but more could be done to validate the results from these tests: Add LD matrix tests, implementing the rest of the summary functions. Mirror the ld_matrix behavior. More thorough testing of various data partitions. Compare to LD calculator, and use all of the example tree sequences from test_highlevel.
- Loading branch information