Meshparty no longer loading

[shandran]

Hello,
I am getting an error now when I attempt to load Meshparty. I was having no issues as recently as a week and half ago when I last ran it. I have copied the error below. My apologies for lack of programming and environment debugging, I'm a biologist with amateur coding skills... :( Any advice is greatly appreciated.

I did upgrade to the latest version of Meshparty today (1.16.7) after getting the error but to no avail though.

Here is my notebook where I now get the error. I'm using Python package from micronsbinder on a Windows laptop. I didn't change anything in the last week and a half (apart from upgrading Meshparty today).

Here is the line that now generates an error when I run it:
from meshparty import trimesh_io, trimesh_vtk, skeleton, utils

---

ValueError Traceback (most recent call last)
~\AppData\Local\Temp\ipykernel_21760\2181983511.py in <cell line: 1>()
----> 1 from meshparty import trimesh_io, trimesh_vtk, skeleton, utils

~\anaconda3\lib\site-packages\meshparty\trimesh_io.py in
28 from trimesh import io as exchange
29
---> 30 from pymeshfix import _meshfix
31 from tqdm import trange
32 import DracoPy

~\anaconda3\lib\site-packages\pymeshfix_init_.py in
----> 1 from pymeshfix.meshfix import *
2 from pymeshfix import _meshfix
3 from pymeshfix._version import version
4 from pymeshfix._meshfix import clean_from_arrays, clean_from_file, PyTMesh

~\anaconda3\lib\site-packages\pymeshfix\meshfix.py in
5 import numpy as np
6
----> 7 from pymeshfix import _meshfix
8
9

pymeshfix\cython_meshfix.pyx in init pymeshfix._meshfix()

ValueError: numpy.ndarray size changed, may indicate binary incompatibility. Expected 96 from C header, got 88 from PyObject

[fcollman]

This is an error which reflects an unfortunate but common problem with python packaging and environments. Your numpy version that is installed does not match the version that a compiled library is expecting. You likely recently installed a new library (unrelated to meshparty or trimesh) which triggered a new version (or old version) of numpy to be installed and caused this. In general this often happens when using a mix of conda and pop to manage your packages, because they don’t know about each others dependancies. I would try upgrading or downgrading numpy in a way that makes those two version numbers 96 and 88 get closer together till they match