Perhaps we could do mutation testing, e.g. with mutmut.
That is, we mutate our (passing) tests automatically, to see whether they then fail. This can be useful because many of our tests do not actually assert much, they just verify that the code runs without raising exceptions. As in, the current tests give us coverage, but no confidence that the result is actually correct. Mutation testing would allow us to see which tests could be improved and how.
Perhaps we could do mutation testing, e.g. with mutmut.
That is, we mutate our (passing) tests automatically, to see whether they then fail. This can be useful because many of our tests do not actually assert much, they just verify that the code runs without raising exceptions. As in, the current tests give us coverage, but no confidence that the result is actually correct. Mutation testing would allow us to see which tests could be improved and how.