[Debian-med-packaging] Bug#1146866: dipy: FTBFS on i386: test failures

[Sebastian Ramacher]

Source: dipy
Version: 1.12.1-5
Severity: serious
Tags: ftbfs
Justification: fails to build from source (but built successfully in the past)
X-Debbugs-Cc: sramacher at debian.org
User: debian-qa at lists.debian.org
Usertags: i386

Dear maintainer,

dipy FTBFS on i386:


=================================== FAILURES ===================================
_______________________________ test_ascm_static _______________________________

    def test_ascm_static():
        S0 = 100 * np.ones((20, 20, 20), dtype="f8")
        S0n1 = nlmeans(S0, sigma=0, rician=False, patch_radius=1, block_radius=1)
        S0n2 = nlmeans(S0, sigma=0, rician=False, patch_radius=2, block_radius=1)
        S0n = adaptive_soft_matching(S0, S0n1, S0n2, 0)
    
>       assert abs(S0n.mean() - 100) < 1
E       assert 1.7775472817319411 < 1
E        +  where 1.7775472817319411 = abs((98.222452718268059 - 100))
E        +    where 98.222452718268059 = <built-in method mean of numpy.ndarray object at 0xe6d87920>()
E        +      where <built-in method mean of numpy.ndarray object at 0xe6d87920> = array([[[ 105.64528906,   83.73337841,  103.14523929, ...,  100.10504015,\n          100.02191643,  100.00381517],\n    ...2],\n        [  97.76005389,   96.61499114,   99.12334288, ...,  100.48079701,\n           99.98568357,  100.01387022]]]).mean

dipy/denoise/tests/test_ascm.py:18: AssertionError
______________________ test_blockwise_sigma_array_support ______________________

    def test_blockwise_sigma_array_support():
        """
        Test that blockwise method supports different sigma input formats.
        """
        S0 = 100 * np.ones((20, 20, 20), dtype="f8")
        result1 = nlmeans(S0, sigma=1.0, method="blockwise")
        assert result1.shape == S0.shape
    
        result2 = nlmeans(S0, sigma=np.array(1.0), method="blockwise")
        assert result2.shape == S0.shape
>       assert np.abs(np.mean(result2) - np.mean(result1)) < 1.0
E       AssertionError: assert 1.1618215408360868 < 1.0
E        +  where 1.1618215408360868 = <ufunc 'absolute'>((98.983997815277846 - 97.822176274441759))
E        +    where <ufunc 'absolute'> = np.abs
E        +    and   98.983997815277846 = <function mean at 0xf5938940>(array([[[ 99.9899995 ,  99.9899995 ,  99.9899995 , ...,  99.9899995 ,\n          99.9899995 ,  99.9899995 ],\n        [ ...899995 ],\n        [ 99.9899995 ,  99.9899995 ,  99.9899995 , ...,  99.9899995 ,\n          99.9899995 ,  99.9899995 ]]]))
E        +      where <function mean at 0xf5938940> = np.mean
E        +    and   97.822176274441759 = <function mean at 0xf5938940>(array([[[ 99.9899995 ,  99.9899995 ,  99.9899995 , ...,  99.9899995 ,\n          99.9899995 ,  99.9899995 ],\n        [ ...899995 ],\n        [ 99.9899995 ,  99.9899995 ,  99.9899995 , ...,  99.9899995 ,\n          99.9899995 ,  99.9899995 ]]]))
E        +      where <function mean at 0xf5938940> = np.mean

dipy/denoise/tests/test_nlmeans.py:199: AssertionError
=========================== short test summary info ============================
FAILED dipy/denoise/tests/test_ascm.py::test_ascm_static - assert 1.777547281...
FAILED dipy/denoise/tests/test_nlmeans.py::test_blockwise_sigma_array_support
==== 2 failed, 957 passed, 174 skipped, 59 deselected in 5727.48s (1:35:27) ====

See
https://buildd.debian.org/status/fetch.php?pkg=dipy&arch=i386&ver=1.12.1-5%2Bb1&stamp=1788589386&raw=0
for details.

Cheers
-- 
Sebastian Ramacher