search

seung-lab / euclidean-distance-transform-3d

Euclidean distance & signed distance transform for multi-label 3D anisotropic images using marching parabolas.
GNU General Public License v3.0
234 stars 37 forks source link

Error importing edt #34

Closed kevinjohncutler closed 3 years ago

kevinjohncutler commented 3 years ago

I am getting the following error when attempting to import edt: numpy.ndarray size changed, may indicate binary incompatibility. Expected 88 from C header, got 80 from PyObject I am using python 3.8.5 on Ubuntu 20.04. Let me know if there is anything else I can do to help debug.

william-silversmith commented 3 years ago

Hi, unfortunately Numpy 1.20 broke binary compatibility. You can either install from source (pip install edt --no-binary :all:), update numpy to 1.20, or install an older version of edt that was compiled against older numpy versions.

Sorry about that... I'm not sure the best way to handle this other than releasing against the most recent numpy.

On Sat, Apr 24, 2021, 4:36 AM Kevin Cutler @.***> wrote:

I am getting the following error when attempting to import edt: numpy.ndarray size changed, may indicate binary incompatibility. Expected 88 from C header, got 80 from PyObject I am using python 3.8.5 on Ubuntu 20.04. Let me know if there is anything else I can do to help debug.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/seung-lab/euclidean-distance-transform-3d/issues/34, or unsubscribe https://github.com/notifications/unsubscribe-auth/AATGQSORNA7YQAIKSJHXYZTTKJ7I5ANCNFSM43P5IYYQ .

william-silversmith commented 3 years ago

Looks like if you compile against oldest-supported-numpy this problem is resolved. I'll have to make a new release so that ABI compatibility is maintained.

william-silversmith commented 3 years ago

Recompiled! Please reopen this issue if it still presents a problem. As a bonus, I was able to release universal2 binaries for py38!