If I remember correctly, the formulations in the planner_biped.py file were faster than the one in the planner_generic.py. However, they seems to use the same problem definition and problemData and they are not compatible anymore. The main issue is that now there is one more list hierarchy level in PhaseData.S and PhaseData.moving.
After trying to solve this API issues, I could not manage to make this problems converge on simple example where the generic planners are working fine.
It fail with "Failed to solve the MIP" or "Initial LP solver fails in fix_sparsity_combinatorial". Even if I make a model with only one candidate surface per phase.
Are this methods still used ? Are they supposed to work in the latest branches ?
This is part of an ongoing problem which is that Sl1m is not tested enough. Before getting rid of it I'd like to have @FannyRis opinion on this although I suspect we won't hear from them before a while
If I remember correctly, the formulations in the planner_biped.py file were faster than the one in the planner_generic.py. However, they seems to use the same problem definition and problemData and they are not compatible anymore. The main issue is that now there is one more list hierarchy level in PhaseData.S and PhaseData.moving.
After trying to solve this API issues, I could not manage to make this problems converge on simple example where the generic planners are working fine. It fail with "Failed to solve the MIP" or "Initial LP solver fails in fix_sparsity_combinatorial". Even if I make a model with only one candidate surface per phase.
Are this methods still used ? Are they supposed to work in the latest branches ?