Add frame visualization to Meshcat visualizer

[sea-bass]

This PR adds frame visualization to the Meshcat visualizer -- the use case here is for some educational content I'm making where I will start with a pure model that has no visual or collision models, and I still want the frames to show up. But I also think it would generally be useful to have this available.

Frame visualization is False by default to not change behavior for users.

The repro case for the animation below looks roughly like this right now:

viz = MeshcatVisualizer(model, collision_model, visual_model, data=data)
viz.initViewer(open=True)
viz.loadViewerModel()
viz.displayFrames(True)  # Default args
viz.displayFrames(True, frame_ids=[10, 11, 12], axis_length=0.2, axis_width=2)  # Optional args
# You can also call viz.initializeFrames() with the same keyword args as above if you want to
# reinitialize frame visualizations with new options, without toggling visibility.

q = pinocchio.randomConfiguration(model)
v = np.zeros_like(q)
tau = np.zeros_like(q)

for i in range(100):
    a1 = pinocchio.aba(model, data, q, v, tau)
    vnext = v + dt * a1
    qnext = pinocchio.integrate(model, q, dt * vnext)

    viz.display(qnext)

    q = qnext
    v = vnext
    time.sleep(dt)

2023-11-25.12-29-47.mp4

[hrp2-14]

Hi ! This project doesn't usually accept pull requests on the main branch.
If this wasn't intentionnal, you can change the base branch of this PR to devel
(No need to close it for that). Best, a bot.

[nim65s]

This looks good to me, thanks for adding this useful feature !

I guess we could add parameters in loadViewerModel to let the user define if they want to display collisions, visuals, and/or frames or not, but this is out of the scope of this PR, and would better be consistently applied on all viewers.

[jcarpent]

Thanks @sea-bass for this very helpful contribution.