Hypergeometric test and function redo #21922
Conversation
|
Thanks for contributing to Ivy! 😊👏 |
ivy-leaves
added
the
NumPy Frontend
| @to_ivy_arrays_and_back | ||
| @from_zero_dim_arrays_to_scalar | ||
| def hypergeometric(ngood, nbad, nsample, size=None): | ||
| u = ivy.empty(size, dtype=int) |
There was a problem hiding this comment.
size argument is default None which is a required argument for empty which will always throw an error. As seen in the test logs here.
There was a problem hiding this comment.
Whenever I reproduce any of the failures I get an error that states that the test passed
There was a problem hiding this comment.
I tested the test cases locally and they don't fail like they do in the test suite. I even compared it to the numpy.random.hypergeometric module and they produce the same results (having in mind that the numbers are different due to the randomness)
There was a problem hiding this comment.
yea makes sense. The issue is a flaw in the logic. If you test call this function and don't specify a value for size then this would break because None is not a valid argument for size in ivy.empty.
There was a problem hiding this comment.
I dont get that issue in the new commits anymore
There was a problem hiding this comment.
Are all tests passing for you locally?
There was a problem hiding this comment.
locally I dont get any errors, yes
There was a problem hiding this comment.
I just want to mention, I have this many commits because my local tests only run for numpy and don't check the other backends.
Close #21387