Import Error

[WYXG233]

Habitat-Sim version

RLRAudioPropagationUpdate & habitat_sim-0.3.1

🐛 Bug

when I try to import habitat_sim.sim, it will output a TypeError. Below is the Traceback.

Steps to Reproduce

1. python
2. import habitat_sim.sim

File "miniconda/envs/ss/lib/python3.9/site-packages/habitat_sim-0.3.1-py3.9-linux-x86_64.egg/habitat_sim/__init__.py", line 14, in <module>
    from habitat_sim import (
  File "miniconda/envs/ss/lib/python3.9/site-packages/habitat_sim-0.3.1-py3.9-linux-x86_64.egg/habitat_sim/agent/__init__.py", line 7, in <module>
    from .agent import *  # noqa: F403
  File "miniconda/envs/ss/lib/python3.9/site-packages/habitat_sim-0.3.1-py3.9-linux-x86_64.egg/habitat_sim/agent/agent.py", line 12, in <module>
    import quaternion as qt
  File "miniconda/envs/ss/lib/python3.9/site-packages/quaternion/__init__.py", line 34, in <module>
    from .calculus import (
  File "miniconda/envs/ss/lib/python3.9/site-packages/quaternion/calculus.py", line 462, in <module>
    from scipy.interpolate import InterpolatedUnivariateSpline
  File "miniconda/envs/ss/lib/python3.9/site-packages/scipy/interpolate/__init__.py", line 167, in <module>
    from ._interpolate import *
  File "miniconda/envs/ss/lib/python3.9/site-packages/scipy/interpolate/_interpolate.py", line 14, in <module>
    from . import _fitpack_py
  File "miniconda/envs/ss/lib/python3.9/site-packages/scipy/interpolate/_fitpack_py.py", line 8, in <module>
    from ._fitpack_impl import bisplrep, bisplev, dblint  # noqa: F401
  File "miniconda/envs/ss/lib/python3.9/site-packages/scipy/interpolate/_fitpack_impl.py", line 103, in <module>
    'iwrk': array([], dfitpack_int), 'u': array([], float),
TypeError

Expected behavior

import correctly

System Info

Platform: Linux-6.5.0-35-generic-x86_64-with-glibc2.35
Machine: x86_64
Processor: x86_64
Libc version: glibc 2.35
Mac version:
Python version: 3.9.19
Architecture: 64bit ELF
Win version:
System OS: Linux
Release: 6.5.0-35-generic
Version: #3522.04.1-Ubuntu SMP PREEMPT_DYNAMIC Tue May 7 09:00:52 UTC 2
Operational System: linux
GCC version: b'gcc (Ubuntu 11.4.0-1ubuntu122.04) 11.4.0\n'
CMAKE version: b'cmake version 3.14.0\n'
NVIDIA-SMI: b"NVIDIA-SMI has failed because it couldn't communicate with the NVIDIA driver. Make sure that the latest NVIDIA driver is installed and running.\n"
Pip packages versions:
b'imageio==2.34.1\nimageio-ffmpeg==0.5.1\nmatplotlib==3.8.4\nnumba==0.59.1\nnumpy==1.26.4\nnumpy-quaternion==2023.0.3\ntqdm==4.66.4\n'
Conda packages versions:
b'attrs 23.2.0 pyh71513ae_0 conda-forge 19.1.0\ngitpython 3.1.43 pyhd8ed1ab_0 conda-forge\nimageio 2.34.1 pyh4b66e23_0 conda-forge\nimageio-ffmpeg 0.5.1 pyhd8ed1ab_0 conda-forge\nmatplotlib 3.8.4 py39hf3d152e_2 conda-forge\nnumba 0.59.1 py39h615d6bd_0 conda-forge\nnumpy 2.0.0 pypi_0 pypi\npillow 8.3.2 py39ha612740_0 conda-forge\nscipy 1.13.1 py39haf93ffa_0 conda-forge 1.3.0\ntqdm 4.66.4 pyhd8ed1ab_0 conda-forge\n'

[ziyuwan]

@WYXG233 have you solved it now? I've also encountered this problem.

[aclegg3]

This looks like a likely dependency version incompatibility between numpy, numpy-quaternion, and numba.

[WYXG233]

@ziyuwan Yeah, I haven't found out how to fix it. But it works correctly in another environment. Maybe you can try the following dependency version:

numpy-1.23.5 (instead of the above numpy 2.0.0)

btw, hope you can update this incompatibility issue with a higher version of numpy.

[ziyuwan]

@WYXG233 Thanks a lot for your reply!
I've also found that my numpy version was wrong. After installing numpy==1.23.0, it works now.

[aclegg3]

For what its worth, my current build is working with:

numba                     0.60.0rc1                pypi_0    pypi
numpy                     1.26.4                   pypi_0    pypi
numpy-quaternion          2023.0.3                 pypi_0    pypi

[samuelm2]

numpy                     1.26.4                   pypi_0    pypi
numpy-base                2.0.1           py310h65a83cf_1

I had a similar issue which happened because of pip and conda having conflicting numpy versions. I must've accidentally used pip to switch numpy versions, but the numpy-base conda package version was still at 2.0.1. So whenever I checked the numpy package version in pip, it was showing 1.26.4. But it still wasn't working with sklearn because I guess something under the hood was still using 2.0.1. I fixed it by doing a conda install numpy-base=1.26.4