Hi all,

the error is apparently caused by a change in SciPy 1.15.0 in the function scipy.spatial.distance.jaccard. Quoting the documentation at https://docs.scipy.org/doc/scipy-1.15.2/reference/generated/scipy.spatial.distance.jaccard.html:

"Changed in version 1.15.0: Non-0/1 numeric input used to produce an ad hoc result. Since 1.15.0, numeric input is converted to Boolean before computation."

I am going to suggest a fix soon. Please stay tuned.

Best,

Daniel


Am 13.04.25 um 8:11 AM schrieb Andreas Tille:
Control: tags -1 upstream
Control: forwarded -1 Daniel Müllner <dan...@danifold.net>
Thanks

Hi Daniel,

as you can see below the Debian packaged version of fastcluster fails
its autopkgtest.  It would be great if you could comment on this issue.
There might be chances that this problem is caused by

   numpy (version 2.2.4)
   scipy (version 1.15.2)

Any hint how to fix the test issue would be welcome.

Kind regards
    Andreas.

The complete information about this bug report can be found here
    https://bugs.debian.org/1100292

Am Wed, Apr 09, 2025 at 09:24:40AM +0900 schrieb Charles Plessy:
Control: tag -1 help

Hi all,

I had a look at the build failure and could reproduce it locally.  Here is
an extract from Santiago's report.

The test failure looks relevant and probably should not ignored.
Unfortunately, the combination of C++ and Python is very, very far from my
expertise.  If somebody else can have a look that would be great.

The alternative is to release without the package and its reverse
dependencies, unless we know that there is no risk to inflict erroneous
computations on our users.

Have a nice day,

Charles

==================================================================================================================
Fastcluster version: 1.2.6
Python version: 3.13.2 (main, Feb  5 2025, 01:23:35) [GCC 14.2.0]
running config
    dh_auto_build -O--buildsystem=pybuild -O--sourcedirectory=src/python
I: pybuild base:311: /usr/bin/python3 setup.py build
/usr/lib/python3/dist-packages/setuptools/_distutils/dist.py:270:
UserWarning: Unknown distribution option: 'test_suite'
   warnings.warn(msg)
Fastcluster version: 1.2.6
Python version: 3.13.2 (main, Feb  5 2025, 01:23:35) [GCC 14.2.0]
running build
running build_py
copying fastcluster.py ->
/<<PKGBUILDDIR>>/.pybuild/cpython3_3.13_fastcluster/build
running build_ext
building '_fastcluster' extension
creating build/temp.linux-x86_64-cpython-313
x86_64-linux-gnu-g++ -g -O2 -ffile-prefix-map=/<<PKGBUILDDIR>>=.
-fstack-protector-strong -fstack-clash-protection -Wformat
-Werror=format-security -fcf-protection -Wdate-time -D_FORTIFY_SOURCE=2
-fPIC -I/usr/lib/x86_64-linux-gnu/python3-numpy/numpy/_core/include
-I/usr/include/python3.13 -c fastcluster_python.cpp -o
build/temp.linux-x86_64-cpython-313/fastcluster_python.o
In file included from fastcluster_python.cpp:38:
../fastcluster.cpp:53:35: note: ‘#pragma message: Use fenv header.’
    53 | #pragma message("Use fenv header.")
       |                                   ^
x86_64-linux-gnu-g++ -g -O2 -ffile-prefix-map=/<<PKGBUILDDIR>>=.
-fstack-protector-strong -fstack-clash-protection -Wformat
-Werror=format-security -fcf-protection -Wdate-time -D_FORTIFY_SOURCE=2
-shared -Wl,-O1 -Wl,-Bsymbolic-functions -Wl,-z,relro -g -fwrapv -O2
-Wl,-z,relro -Wl,-z,now -g -O2 -ffile-prefix-map=/<<PKGBUILDDIR>>=.
-fstack-protector-strong -fstack-clash-protection -Wformat
-Werror=format-security -fcf-protection -Wdate-time -D_FORTIFY_SOURCE=2
build/temp.linux-x86_64-cpython-313/fastcluster_python.o
-L/usr/lib/x86_64-linux-gnu -o 
/<<PKGBUILDDIR>>/.pybuild/cpython3_3.13_fastcluster/build/_fastcluster.cpython-313-x86_64-linux-gnu.so
    dh_auto_test -O--buildsystem=pybuild -O--sourcedirectory=src/python
I: pybuild base:311: cd
/<<PKGBUILDDIR>>/.pybuild/cpython3_3.13_fastcluster/build; python3.13 -m
unittest discover -v
test (tests.fastcluster_test.test) ... ok
test_nan (tests.fastcluster_test.test_nan) ... ok
test_vector (tests.fastcluster_test.test_vector) ... FAIL

======================================================================
FAIL: test_vector (tests.fastcluster_test.test_vector)
----------------------------------------------------------------------
Traceback (most recent call last):
   File 
"/<<PKGBUILDDIR>>/.pybuild/cpython3_3.13_fastcluster/build/tests/__init__.py",
line 14, in test_vector
     test(10)
     ~~~~^^^^
   File 
"/<<PKGBUILDDIR>>/.pybuild/cpython3_3.13_fastcluster/build/tests/vectortest.py",
line 236, in test
     test_all(n,dim)
     ~~~~~~~~^^^^^^^
   File 
"/<<PKGBUILDDIR>>/.pybuild/cpython3_3.13_fastcluster/build/tests/vectortest.py",
line 115, in test_all
     check(Z2, method, D)
     ~~~~~^^^^^^^^^^^^^^^
   File 
"/<<PKGBUILDDIR>>/.pybuild/cpython3_3.13_fastcluster/build/tests/vectortest.py",
line 142, in check
     raise AssertionError(
         'Not the global minimum in step {2}: {0}, {1}'.
         format(Z2[i,2], gmin,i), squareform(D))
AssertionError: ('Not the global minimum in step 0: 0.42857142857142855,
0.0', array([[0.        , 0.28571429, 0.14285714, ..., 0.33333333,
0.14285714,
         0.14285714],
        [0.28571429, 0.        , 0.14285714, ..., 0.57142857, 0.14285714,
         0.14285714],
        [0.14285714, 0.14285714, 0.        , ..., 0.42857143, 0.        ,
         0.        ],
        ...,
        [0.33333333, 0.57142857, 0.42857143, ..., 0.        , 0.42857143,
         0.42857143],
        [0.14285714, 0.14285714, 0.        , ..., 0.42857143, 0.        ,
         0.        ],
        [0.14285714, 0.14285714, 0.        , ..., 0.42857143, 0.        ,
         0.        ]], shape=(62, 62)))

----------------------------------------------------------------------
Ran 3 tests in 1.095s

==================================================================================================================

--
Charles Plessy - - ~ ~ ~ ~ ~ ~~~~ ~ ~ ~ ~ ~ - - charles.ple...@oist.jp
Okinawa  Institute  of  Science  and  Technology  Graduate  University
Senior staff scientist ~~ Luscombe Unit ~~ https://groups.oist.jp/grsu
Toots from work - ~~ ~~~~ ~~ - https://fediscience.org/@charles_plessy



Reply via email to