bipedal-locomotion-framework: ImportError: arg(): could not convert default argument into a Python object when using BLF
Hi @traversaro @S-Dafarra @GiulioRomualdi ,
at very beginning when importing BLF python module, I had some other issues which are reported here: https://github.com/ami-iit/element_pi-learning-human-estimation/issues/36#issuecomment-1876906558, this morning I thought I finally fixed the problem, but then I found I can’t use TomlParametersHandler()
from bipedal_locomotion_framework.bindings.parameters_handler
. So I did some uninstall-cleaning-install things, then I found myself stuck in the following bug if I try to import the BLF module:
Traceback (most recent call last):
File "IKSolver.py", line 4, in <module>
import bipedal_locomotion_framework.bindings.parameters_handler as blfh
File "/home/cheng/robotology-superbuild/build/install/lib/python3/dist-packages/bipedal_locomotion_framework/__init__.py", line 1, in <module>
from .bindings import *
ImportError: arg(): could not convert default argument into a Python object (type not registered yet?). Compile in debug mode for more information.
I’m using a virtual environment with python version 3.8.10. Before importing BLF python module, I did following things:
- In
robotology/src/manif
switch todevel
branch and update to the newest, then inrobotology/build/src/manif
I didcmake
+make install
. - In
robotology/src/blf
switch tov0.17.0
, then inrobotology/build/src/blf
I didcmake
+make install
things. - Then I did
pip install git+https://github.com/artivis/manif@devel
.
In the end I have bipedal_locomotion_framework: 0.17.0
and manifpy: 0.1.dev756
installed. I can import manifpy
without any problem. I tried to find the reason of above mentioned issue with BLF importing, and only got this: https://github.com/patrikhuber/eos/issues/256 and this: https://github.com/neumond/pybrood/issues/4, which are most related to my case.
Do you have any idea what could be the reason and how could I tackle it? Many thanks 😃
About this issue
- Original URL
- State: closed
- Created 6 months ago
- Comments: 18 (2 by maintainers)
In the end thanks to @traversaro, the problem is solved by creating a clean conda env instead of mix-using multiple package managers. I think we can close this issue now.
That’s exactly what I was about to ask. I’d love to do that!