Display issues with `hector` drone in pin3

[fabinsch]

Since I use pinocchio3 from conda, the hector drone from example-robot-data is not displayed properly in meshcat. This is not blocking me, I just thought I raise the issue as maybe others will experience display issues as well. I displayed a couple other quadrupeds and manipulator models from erd and they all look fine.

Steps to reproduce:

conda create --name erd_debug python=3.11 pinocchio example-robot-data meshcat-python ipython

then in ipython

import example_robot_data as erd
from pinocchio.visualize import MeshcatVisualizer
robot = erd.load("hector")
vizer =  MeshcatVisualizer(
    robot.model, robot.collision_model, robot.visual_model
)
vizer.initViewer(loadModel=True, open=True)
vizer.display(robot.q0)

It works as expected with

conda create --name erd_debug python=3.11 pinocchio==2.7.1 example-robot-data meshcat-python ipython

[jorisv]

I think @ManifoldFR have found a similar issue with the UR5 end effector.

[ManifoldFR]

What is the file extension for the meshes for this robot ?

If it's STL or DAE I have a fix over on our own fork of meshcat-python on the Simple-Robotics org

Probably related to my PR: https://github.com/meshcat-dev/meshcat/pull/170

[fabinsch]

You are right, the UR5 end-effector has also display issues. However, they look different than what we see on your PR @ManifoldFR .

The drone has both DAE for visual and STL for collisions. I installed the latest master from our fork of meshcat-python but this does not solve the problem.

[ManifoldFR]

Oh, the branch to install is not master, but topic/update-meshcatjs I think.

The current master on the fork is just the master from upstream if I recall correctly.

[ManifoldFR]

That one should point to the right branch on our fork of meshcat (JS) which has all of the patches I wanted to upstream. Unfortunately nothing has advanced on that front.

[fabinsch]

mmh, I tried also the branch topic/update-meshcatjs with commit 819526d46f9c63efad7a0f3df9a15db0cc2ce0c9 but it looks still the same.

[fabinsch]

The issue is solved for me by #2294.