Closed cielavenir closed 3 months ago
sometimes mpmath.polyroots does not converge well, so I relax the threshold.
lib/python3.9/site-packages/mpmath/calculus/polynomials.py in polyroots(ctx, coeffs, maxsteps, cleanup, extraprec, error) 185 err[i] = abs(x) 186 if abs(max(err)) >= tol: --> 187 raise ctx.NoConvergence("Didn't converge in maxsteps=%d steps." \ 188 % maxsteps) 189 # Remove small real or imaginary parts NoConvergence: Didn't converge in maxsteps=50 steps.
with this patch, I locally confirm that m____c50h iksolver can be generated with sympy 1.11.1. Also I checked some ikparams can be solved.
cc @ntohge
as side-effect, https://github.com/roboticsleeds/ur5controller/issues/7 and https://github.com/andyzeng/ikfastpy/issues/2 and https://github.com/rdiankov/openrave/issues/578 will be solved positively (if threshold is alright).
Thanks~
sometimes mpmath.polyroots does not converge well, so I relax the threshold.
with this patch, I locally confirm that m____c50h iksolver can be generated with sympy 1.11.1. Also I checked some ikparams can be solved.
cc @ntohge
as side-effect, https://github.com/roboticsleeds/ur5controller/issues/7 and https://github.com/andyzeng/ikfastpy/issues/2 and https://github.com/rdiankov/openrave/issues/578 will be solved positively (if threshold is alright).