Jason K. Moore

With commit ef6a9e7 it's showing that the C code generated is not giving machine precision match on the evaluation of the mass matrix and forcing vector. The follow example shows...

It now seems clear that the floating point errors are causing the larger differences with the Basu-Mandal numbers. Using symengine to evaluate the expressions at an exact precision gives the...

I solved for the udots symbolically (LUSolve of the 6x6 matrix) and then CythonODEFunctionGenerated that. It took like 12+ hours to compile, which is nutz. I am not sure why...

We work somewhat on the "garbage in/garbage out" principal instead of having tons of type checks, but if you'd like to submit a PR with the type checks I will...

This would be a good one to start with: https://github.com/pydy/pydy/issues/296 You have to run the examples and see which ones are using deprecated code.

Note that the shapes support sympy expressions as dimensions, so they shouldn't "only be int or float". I actually don't see any reason to make these specific type checks. Python...

This is going to need to be in a directory for this particular example. What is the example? Is a two mass chain link?

@sahilshekhawat I think most of these examples really need to be unit tests. You can make examples and unit tests, but these super basic examples should be used as tests,...

Thanks for the issue. Yes, these tests have been problematic. I may have an open issue about converting them to only numerical checks. It will take some effort to do...

@tarzzz This is close to working now. I think I must just have a small bug. I'll play with it more later. But you can see if you can find...