Creating a minimal example to tackle 'ValueError: The truth value of an array with more than one element is ambiguous' when building JaxSimModelData instance

[lorycontixd]

This issue starts from my robot implementation which uses code from #156 , where building a JaxSimModelData instance inside a loop raises the following error:

Traceback (most recent call last):
File "/mnt/c/Users/lconti/Documents/programming/dev/element_lower-leg-morphology-optimization/src/passive_walker/3-Learner/learner_grid.py", line 661, in
main()
File "/mnt/c/Users/lconti/Documents/programming/dev/element_lower-leg-morphology-optimization/src/passive_walker/3-Learner/learner_grid.py", line 619, in main
soft_contacts_params=js.contact.estimate_good_soft_contacts_parameters(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "", line 4, in eq
File "/home/lorenzoconti/miniforge3/envs/pw_learning/lib/python3.12/site-packages/jax/_src/array.py", line 258, in bool
core.check_bool_conversion(self)
File "/home/lorenzoconti/miniforge3/envs/pw_learning/lib/python3.12/site-packages/jax/_src/core.py", line 654, in check_bool_conversion
raise ValueError("The truth value of an array with more than one element"
ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all()

As of now, I tried recreating a minimal working example that attempts to recreate the problem, but this example runs correctly.

Code

Code link here.

Code file (txt): minimal.txt

Conda env

requirements.txt

File diff

diff.txt

Next step is to add more lines from the original file to try to break the code with the traceback found above.

CC: @flferretti

Tasks

Preview Give feedback

1. Check if joints break the data construction
2. Add initial force to joints under data's link_forces parameter
3. Check if removed custom frames break the code
4. Check if frame convention switch breaks the code

[flferretti]

Hi @lorycontixd, can you please post the result from diff -u working_version.py non_working_version.py? Thanks a lot

[lorycontixd]

Hi @lorycontixd, can you please post the result from diff -u working_version.py non_working_version.py? Thanks a lot

I updated the original comment with the output of the diff. I have a feeling the code block has a length limit so I uploaded a file for it.

[lorycontixd]

Added joint initial positions & link forces

• Code here
• Diff here

Result: still no breaking of the code

[lorycontixd]

Added frame convention switch:

• Code here
• Diff here

Result: still no breaking of the code

[diegoferigo]

@lorycontixd can you please add snippets of MRE in code boxes wrapped by a details HTML environment? The repositories you linked are mostly private and people reading this issue have no access.

[lorycontixd]

@lorycontixd can you please add snippets of MRE in code boxes wrapped by a details HTML environment? The repositories you linked are mostly private and people reading this issue have no access.

Sure, do code boxes have a maximum limit? Because the code is 400+ lines and most of the times I have difficulties wrapping code in the appropriate boxes.

[diegoferigo]

There are no significant limits AFAIK. However, in order to help other people providing you support, the shorter the MRE is, the better.

[lorycontixd]

I updated the original comment to state that the code makes use of #156 @diegoferigo

[flferretti]

Closed by #177