<div dir="ltr"><br><div class="gmail_extra"><br><br><div class="gmail_quote">On Mon, Sep 30, 2013 at 5:12 PM, Christoph Gohlke <span dir="ltr"><<a href="mailto:cgohlke@uci.edu" target="_blank">cgohlke@uci.edu</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="im">On 9/30/2013 3:45 PM, Charles R Harris wrote:<br>
><br>
><br>
><br>
> On Mon, Sep 30, 2013 at 3:51 PM, Christoph Gohlke <<a href="mailto:cgohlke@uci.edu">cgohlke@uci.edu</a><br>
</div><div class="im">> <mailto:<a href="mailto:cgohlke@uci.edu">cgohlke@uci.edu</a>>> wrote:<br>
><br>
><br>
><br>
> On 9/30/2013 11:02 AM, Nathaniel Smith wrote:> Everyone please do<br>
> actually test this! It is really in your best<br>
> > interest, and I think people don't always realize this.<br>
> ><br>
> > Here's how it works:<br>
> > - If you test it *now*, and it breaks your code that worked with 1.7,<br>
> > and you *tell* us this now, then it's *our* problem and we hold up the<br>
> > release to fix the bug.<br>
> > - If you test it *after* we release, and it breaks your code, then we<br>
> > are sad but you have to work around it (because we can't magically<br>
> > make that release not have happened, your users will be using it<br>
> > anyway), and we put it on the stack with all the other bugs. All of<br>
> > which we care about but it's a large enough stack that it's not going<br>
> > to get any special priority, because, see above about how at this<br>
> > point you'll have had to work around it anyway.<br>
> ><br>
> > -n<br>
> ><br>
> > On Mon, Sep 30, 2013 at 4:17 PM, Charles R Harris<br>
</div><div><div class="h5">> > <<a href="mailto:charlesr.harris@gmail.com">charlesr.harris@gmail.com</a> <mailto:<a href="mailto:charlesr.harris@gmail.com">charlesr.harris@gmail.com</a>>> wrote:<br>
> >> Hi All,<br>
> >><br>
> >> NumPy 1.8.0rc1 is up now on sourceforge .The binary builds are included<br>
> >> except for Python 3.3 on windows, which will arrive later. Many thanks to<br>
> >> Ralf for the binaries, and to those who found and fixed the bugs in the last<br>
> >> beta. Any remaining bugs are all my fault ;) I hope this will be the last<br>
> >> release before final, so please test it thoroughly.<br>
> >><br>
> >> Chuck<br>
><br>
><br>
> Hello,<br>
><br>
> NumPy 1.8.0rc1 looks good. All tests pass on Windows and most 3rd party<br>
> packages test OK now. Thank you.<br>
><br>
> A few tests still fail in the following packages when run with<br>
> numpy-MKL-1.8.0rc1-win-amd64-py3.3 compared to<br>
> numpy-MKL-1.7.1-win-amd64-py3.3:<br>
><br>
> 1) Pandas 0.12.0<br>
><br>
> ```<br>
> ======================================================================<br>
> FAIL: test_nansum_buglet (pandas.tests.test_series.TestNanops)<br>
> ----------------------------------------------------------------------<br>
> Traceback (most recent call last):<br>
> File "X:\Python33\lib\site-packages\pandas\tests\test_series.py",<br>
> line 254, in test_nansum_buglet<br>
> assert_almost_equal(result, 1)<br>
> File "X:\Python33\lib\site-packages\pandas\util\testing.py", line<br>
> 134, in assert_almost_equal<br>
> np.testing.assert_(isiterable(b))<br>
> File "D:\Dev\Compile\Test\numpy-build\numpy\testing\utils.py", line<br>
> 44, in assert_<br>
> raise AssertionError(msg)<br>
> AssertionError<br>
> ```<br>
><br>
> Possibly related:<br>
><br>
> ```<br>
> >>> import numpy as np<br>
> >>> from pandas import Series<br>
> >>> s = Series([0.0])<br>
> >>> result = np.nansum(s)<br>
> >>> print(result)<br>
> Traceback (most recent call last):<br>
> File "<stdin>", line 1, in <module><br>
> File "X:\Python33\lib\site-packages\pandas\core\base.py", line<br>
> 19, in<br>
> __str__<br>
> return self.__unicode__()<br>
> File "X:\Python33\lib\site-packages\pandas\core\series.py", line<br>
> 1115, in __unicode__<br>
> length=len(self) > 50,<br>
> TypeError: len() of unsized object<br>
> ```<br>
><br>
> 2) Bottleneck 0.7.0<br>
><br>
> <a href="https://github.com/kwgoodman/bottleneck/issues/71#issuecomment-25331701" target="_blank">https://github.com/kwgoodman/bottleneck/issues/71#issuecomment-25331701</a><br>
><br>
> 3) skimage 0.8.2<br>
><br>
> These tests passed with numpy 1.8.0b2:<br>
><br>
> ```<br>
> ======================================================================<br>
> FAIL: test_grey.test_non_square_image<br>
> ----------------------------------------------------------------------<br>
> Traceback (most recent call last):<br>
> File "X:\Python33\lib\site-packages\nose\case.py", line 198, in<br>
> runTest<br>
> self.test(*self.arg)<br>
> File<br>
> "X:\Python33\lib\site-packages\skimage\morphology\tests\test_grey.py",<br>
> line 162, in test_non_square_image<br>
> testing.assert_array_equal(binary_res, grey_res)<br>
> File "X:\Python33\lib\site-packages\numpy\testing\utils.py", line<br>
> 718, in assert_array_equal<br>
> verbose=verbose, header='Arrays are not equal')<br>
> File "X:\Python33\lib\site-packages\numpy\testing\utils.py", line<br>
> 644, in assert_array_compare<br>
> raise AssertionError(msg)<br>
> AssertionError:<br>
> Arrays are not equal<br>
><br>
> (mismatch 50.6328125%)<br>
> x: array([[False, False, False, ..., False, False, False],<br>
> [False, False, False, ..., False, False, False],<br>
> [False, False, False, ..., False, False, False],...<br>
> y: array([[ True, True, True, ..., True, False, False],<br>
> [ True, True, True, ..., False, False, False],<br>
> [ True, True, True, ..., False, False, False],...<br>
><br>
> ======================================================================<br>
> FAIL: test_grey.test_binary_erosion<br>
> ----------------------------------------------------------------------<br>
> Traceback (most recent call last):<br>
> File "X:\Python33\lib\site-packages\nose\case.py", line 198, in<br>
> runTest<br>
> self.test(*self.arg)<br>
> File<br>
> "X:\Python33\lib\site-packages\skimage\morphology\tests\test_grey.py",<br>
> line 169, in test_binary_erosion<br>
> testing.assert_array_equal(binary_res, grey_res)<br>
> File "X:\Python33\lib\site-packages\numpy\testing\utils.py", line<br>
> 718, in assert_array_equal<br>
> verbose=verbose, header='Arrays are not equal')<br>
> File "X:\Python33\lib\site-packages\numpy\testing\utils.py", line<br>
> 644, in assert_array_compare<br>
> raise AssertionError(msg)<br>
> AssertionError:<br>
> Arrays are not equal<br>
><br>
> (mismatch 48.260498046875%)<br>
> x: array([[False, False, False, ..., False, False, False],<br>
> [False, False, False, ..., False, False, False],<br>
> [False, False, False, ..., False, False, False],...<br>
> y: array([[ True, True, True, ..., True, False, False],<br>
> [ True, True, True, ..., False, False, False],<br>
> [ True, True, True, ..., False, False, False],...<br>
><br>
> ======================================================================<br>
> FAIL: test_grey.test_binary_closing<br>
> ----------------------------------------------------------------------<br>
> Traceback (most recent call last):<br>
> File "X:\Python33\lib\site-packages\nose\case.py", line 198, in<br>
> runTest<br>
> self.test(*self.arg)<br>
> File<br>
> "X:\Python33\lib\site-packages\skimage\morphology\tests\test_grey.py",<br>
> line 183, in test_binary_closing<br>
> testing.assert_array_equal(binary_res, grey_res)<br>
> File "X:\Python33\lib\site-packages\numpy\testing\utils.py", line<br>
> 718, in assert_array_equal<br>
> verbose=verbose, header='Arrays are not equal')<br>
> File "X:\Python33\lib\site-packages\numpy\testing\utils.py", line<br>
> 644, in assert_array_compare<br>
> raise AssertionError(msg)<br>
> AssertionError:<br>
> Arrays are not equal<br>
><br>
> (mismatch 66.302490234375%)<br>
> x: array([[False, False, False, ..., False, False, False],<br>
> [False, False, False, ..., False, False, False],<br>
> [False, False, False, ..., False, False, False],...<br>
> y: array([[ True, True, True, ..., True, True, True],<br>
> [ True, True, True, ..., True, True, True],<br>
> [ True, True, True, ..., False, False, False],...<br>
><br>
> ======================================================================<br>
> FAIL: test_grey.test_binary_opening<br>
> ----------------------------------------------------------------------<br>
> Traceback (most recent call last):<br>
> File "X:\Python33\lib\site-packages\nose\case.py", line 198, in<br>
> runTest<br>
> self.test(*self.arg)<br>
> File<br>
> "X:\Python33\lib\site-packages\skimage\morphology\tests\test_grey.py",<br>
> line 190, in test_binary_opening<br>
> testing.assert_array_equal(binary_res, grey_res)<br>
> File "X:\Python33\lib\site-packages\numpy\testing\utils.py", line<br>
> 718, in assert_array_equal<br>
> verbose=verbose, header='Arrays are not equal')<br>
> File "X:\Python33\lib\site-packages\numpy\testing\utils.py", line<br>
> 644, in assert_array_compare<br>
> raise AssertionError(msg)<br>
> AssertionError:<br>
> Arrays are not equal<br>
><br>
> (mismatch 58.465576171875%)<br>
> x: array([[False, False, False, ..., False, False, False],<br>
> [False, False, False, ..., False, False, False],<br>
> [False, False, False, ..., False, False, False],...<br>
> y: array([[ True, True, True, ..., True, True, False],<br>
> [ True, True, True, ..., True, True, False],<br>
> [ True, True, True, ..., False, False, False],...<br>
> ```<br>
><br>
> I'll bet the skimage problems come from<br>
> <a href="https://github.com/numpy/numpy/pull/3811" target="_blank">https://github.com/numpy/numpy/pull/3811</a>. They may be doing something<br>
> naughty...<br>
><br>
> Chuck<br>
><br>
<br>
</div></div>A bool image is convolved with a uint8 kernel and the result compared<br>
for equality with an uint32 scalar...<br>
<br>
<a href="https://github.com/scikit-image/scikit-image/blob/master/skimage/morphology/binary.py#L32" target="_blank">https://github.com/scikit-image/scikit-image/blob/master/skimage/morphology/binary.py#L32</a><br>
<div class="HOEnZb"><div class="h5"><br></div></div></blockquote><div><br></div><div>Looks like the result of the convolution is probably output as a bool, which now means 0,1, which does not work when checking equality with the number of pixels in the kernel. I'd call expressing the result of a convolution as a boolean very naughty.<br>
<br></div><div>Chuck <br></div><br></div></div></div>