Issue with rigid_contact_forces plots in multi-phase simulation (Phase 2 curves empty and titles duplicated)

[rdavaill]

I am encountering an issue with the visualization of rigid_contact_forces plots in a multi-phase walking simulation with two consecutive phases using different contacts. Initially, I had overlapping plots because the contact names were identical between phases. I fixed this by assigning unique names to the contacts for each phase (e.g., RHEE_contact1 and RHEE_contact2). Now, the correct number of subplots is generated, but a new issue occurs: the curves for phase 2 do not display at all (the plots remain empty), and the subplot titles from Phase 1 are being duplicated over the Phase 2 plots. Has anyone faced this before, or is there a specific way to force the plotting function to properly link Phase 2 data to its respective axes? Any help would be greatly appreciated!

Debug_Graphs.py

PAT0_scaled(DoubleLimbL2RPhase1).txt

PAT0_scaled(SwingDroit).txt

[Ipuch]

Can you generate an example with 2 contact points only ?

For example,

• a cube with 2 dof xy translations
• Two phases

This would be easier to debug. And tell me if it reproduce the bug.

[rdavaill]

cube_0.txt
cube_1.txt

Debug_Graphs.py

The plotting issue occurs only when the very first contact point (Index 0) of Phase 1 is removed in Phase 2. If a non-deleted point remains at Index 0, it works, but only for a simple deletion.

If a contact point is removed and a new one is created simultaneously during the phase transition, the workaround fails and the bug occurs.

[mickaelbegon]

A fix is available in draft PR #1062. The root cause was the use of global contact indices to address phase-local rigid-contact force rows. The regression case where the first global contact is removed in phase 2 is covered, and the targeted test module passes (7 passed). @pariterre, the implementation is working locally and is ready for review.