`actor_from_primitive` applies double position offset when using multiple centers

[faris-agour]

With actor_from_primitive(..., centers=[...]), the geometry is already translated to each center by repeat_primitive.
If the actor is also shifted by centers[0], everything gets translated twice (double offset).
As a result, the repeated actor does not align with the expected positions (markers / individual actors).

Here a screenshot of the issue

Reproduction code for the issue

import numpy as np

import fury.primitive as fp
from fury.actor.core import actor_from_primitive
from fury import actor, window

centers = np.array([[3, 0, 0], [6, 0, 0], [9, 0, 0]], dtype=np.float32)
colors = np.array([[1, 0, 0], [0, 1, 0], [0, 0, 1]])

vertices, faces = fp.prim_sphere(phi=16, theta=16)


multi = actor_from_primitive(
    vertices, faces, centers, colors=colors, scales=1.0, smooth=True
)



singles = []
for i in range(3):
    s = actor_from_primitive(
        vertices, faces, centers[i : i + 1],
        colors=colors[i : i + 1], scales=1.0, smooth=True,
    )
    s.local.y -= 4
    singles.append(s)


marker_centers = np.vstack([centers, centers - [0, 4, 0]])
markers = actor.sphere(
    centers=marker_centers,
    colors=(1, 1, 1),
    radii=0.15,
)


label_top = actor.text(
    "repeat_primitive (1 actor)",
    position=[-2, 0, 0],
    colors=(1, 1, 1),
    anchor="middle-right",
)
label_bot = actor.text(
    "individual actors",
    position=[-2, -4, 0],
    colors=(1, 1, 1),
    anchor="middle-right",
)


scene = window.Scene()
scene.background = (0.1, 0.1, 0.15)

scene.add(multi)
for s in singles:
    scene.add(s)
scene.add(markers)
scene.add(label_top)
scene.add(label_bot)

show_m = window.ShowManager(
    scene=scene, size=(900, 500), title="repeat_primitive offset test"
)
show_m.start()

[maharshi-gor]

Hi @faris-agour Thanks for reporting this and I see a PR is already open. Let me know once you have made some progress.

[faris-agour]

Hi @maharshi-gor I’ve pushed the requested changes in PR #1125, and it’s now ready for review.