partially initialized module 'ipyvolume' has no attribute '_version' (most likely due to a circular import)

[banxia688]

I have no idea with this problem

[T0bC]

Same Issue:

Used this conda call to create the environment on a WSL Ubuntu:

conda create --solver=libmamba -n rapids-23.12 -c rapidsai -c conda-forge -c nvidia \ rapids=23.12 python=3.10 cuda-version=11.8 \ jupyterlab napari dask-cuda dask-image matplotlib scikit-image anaconda::notebook

then

pip install ipyvolume

import ipyvolume as ipv import numpy as np x, y, z, u, v, w = np.random.random((6, 1000))*2-1 ipv.quickquiver(x, y, z, u, v, w, size=5)

`--------------------------------------------------------------------------- AttributeError Traceback (most recent call last) Cell In[10], line 1 ----> 1 import ipyvolume as ipv 2 import numpy as np 3 x, y, z, u, v, w = np.random.random((6, 1000))*2-1

File ~/miniconda3/envs/rapids-23.12/lib/python3.10/site-packages/ipyvolume/init.py:8 6 from ipyvolume import datasets # noqa: F401 7 from ipyvolume import embed # noqa: F401 ----> 8 from ipyvolume.widgets import # noqa: F401, F403 9 from ipyvolume.transferfunction import # noqa: F401, F403 10 from ipyvolume.pylab import * # noqa: F401, F403

File ~/miniconda3/envs/rapids-23.12/lib/python3.10/site-packages/ipyvolume/widgets.py:30 22 from ipyvolume.traittypes import Image 23 from ipyvolume.serialize import ( 24 array_cube_tile_serialization, 25 array_serialization, (...) 28 texture_serialization, 29 ) ---> 30 from ipyvolume.transferfunction import TransferFunction 31 from ipyvolume.utils import debounced, grid_slice, reduce_size 34 _last_figure = None

File ~/miniconda3/envs/rapids-23.12/lib/python3.10/site-packages/ipyvolume/transferfunction.py:19 17 N = 1024 18 x = np.linspace(0, 1, N, endpoint=True) ---> 19 semver_range_frontend = "~" + ipyvolume._version.__version_js__ 22 @widgets.register 23 class TransferFunction(widgets.DOMWidget): 24 _model_name = Unicode('TransferFunctionModel').tag(sync=True)

AttributeError: partially initialized module 'ipyvolume' has no attribute '_version' (most likely due to a circular import)`

[palec87]

1. Unistalling other widgets such as plotly, k3d etc.
2. Restart the kernel.

Import then worked, unfortunately the docs examples still do no show any plots. @T0bC, did it work for you in the end?

[T0bC]

@palec87

1. Unistalling other widgets such as plotly, k3d etc.
2. Restart the kernel.

Import then worked, unfortunately the docs examples still do no show any plots. @T0bC, did it work for you in the end?

I tried the same thing, but without success. I ended up using a different library. But I must admit that I'm still looking for a python library capable of displaying a 3D array of potentially 1000x1000x1000 voxels with alpha transparency. And the whole thing interactively in real time. I want to display a uCT dataset for segmentation processes without having to reduce the resolution. The software CTVox provided by the manufacturer seems to be able to do this, but unfortunately without pyhton integration.

[palec87]

There you might run into hardware limitation and GPU would help a lot. Here is python CT framework with GPU support https://fast.eriksmistad.no/python-tutorial-mri-ct.html

Other, general use viewer to consider can be napari https://napari.org/stable/

I wanted zero install demos for students, so that it runs in notebooks/colabs. k3d worked for me in the end, but it is quite resource intensive, especially crashing colab NBs.