Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Flaky Test when seed is removed #34

Open
crawlingcub opened this issue Oct 4, 2021 · 0 comments
Open

Flaky Test when seed is removed #34

crawlingcub opened this issue Oct 4, 2021 · 0 comments

Comments

@crawlingcub
Copy link

The test test_boptim in test/test_boptim.py is flaky when the seed is removed. The test failed 500 out of 500 times that I ran.

Interestingly, the test always only fails for the first parameter ("ei") and passes for the other two: "poi" and "cb". I observed that this is true even if I change the order of the parameters i.e., whatever comes first fails! Error message below.

Do you know why this happens or how can this be resolved? I will be happy to raise a PR if you have any suggestions!

Thanks!

Error message:

_________________________________________________________ test_boptim[ei-/home/saikat/projects/borntobeflaky/projects/GPim/test/test_data/test_ei.npy] _________________________________________________________

acqf = 'ei', result = '/home/saikat/projects/borntobeflaky/projects/GPim/test/test_data/test_ei.npy'

    @pytest.mark.parametrize(
        "acqf, result",
        [("ei", test_img_ei),
    #     ("poi", test_img_poi),
    #     ("cb", test_img_cb)
        ])
    def test_boptim(acqf, result):
        Z_sparse = initial_seed()
        X_full = gprutils.get_full_grid(Z_sparse)
        X_sparse = gprutils.get_sparse_grid(Z_sparse)
        expected_result = np.load(result)
        boptim = boptimizer(
            X_sparse, Z_sparse, X_full,
            trial_func, acquisition_function=acqf,
            exploration_steps=20,
            use_gpu=False, verbose=1)
        boptim.run()
        #print(expected_result)
>       assert_allclose(boptim.target_func_vals[-1], expected_result)

test/test_boptim.py:60:
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _  
x = array([[3.69114071e-08,            nan,            nan,            nan,
                   nan,            nan,       ... nan,            nan,
                   nan,            nan,            nan,            nan,
        7.57299443e-34]])
y = array([[3.69114071e-08,            nan,            nan,            nan,
                   nan,            nan, 3.6477... nan,            nan,
                   nan,            nan,            nan,            nan,
        7.57299443e-34]])

func = <ufunc 'isnan'>, hasval = 'nan'

    def func_assert_same_pos(x, y, func=isnan, hasval='nan'):
        """Handling nan/inf.

        Combine results of running func on x and y, checking that they are True
        at the same locations.

        """
        x_id = func(x)
        y_id = func(y)
        # We include work-arounds here to handle three types of slightly
        # pathological ndarray subclasses:
        # (1) all() on `masked` array scalars can return masked arrays, so we
        #     use != True
        # (2) __eq__ on some ndarray subclasses returns Python booleans
        #     instead of element-wise comparisons, so we cast to bool_() and
        #     use isinstance(..., bool) checks
        # (3) subclasses with bare-bones __array_function__ implementations may
        #     not implement np.all(), so favor using the .all() method
        # We are not committed to supporting such subclasses, but it's nice to
        # support them if possible.
        if bool_(x_id == y_id).all() != True:                                                                                                                                                                                msg = build_err_msg([x, y],
                                err_msg + '\nx and y %s location mismatch:'
                                % (hasval), verbose=verbose, header=header,
                                names=('x', 'y'), precision=precision)
>           raise AssertionError(msg)
E           AssertionError:
E           Not equal to tolerance rtol=1e-07, atol=0
E
E           x and y nan location mismatch:
E            x: array([[3.691141e-08,          nan,          nan,          nan,
E                            nan,          nan,          nan,          nan,
E                   1.886172e-02,          nan,          nan, 2.844261e-02,...
E            y: array([[3.691141e-08,          nan,          nan,          nan,
E                            nan,          nan, 3.647780e-03,          nan,
E                            nan, 2.844261e-02,          nan,          nan,...

../../../../anaconda3/envs/GPim/lib/python3.6/site-packages/numpy/testing/_private/utils.py:740: AssertionError
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant