The following will not do what you'd expect, since it doesn't know that the main component has multiple instances:
import rerun as rr
rr.init("rerun_example_points3d_simple", spawn=True)
rr.set_time_sequence("frame", 1)
rr.log("points", rr.Points3D([[0, 0, 0], [1, 1, 1]]))
# This will only color `point[0]`.
rr.set_time_sequence("frame", 2)
rr.log_components("points", [rr.components.ColorBatch([0xFF0000FF])])Worse, there's no public way to explicitly indicate that we want it to splat.
The only way is to go through multiple layers of private APIs:
rr.log_components("points", [rr.components.ColorBatch([0xFF0000FF]), rr._log._splat()])and even then, the above does not work, you need to recast it some more for some reason:
rr.log_components("points", [rr.components.ColorBatch([0xFF0000FF]), rr.components.InstanceKeyBatch(rr._log._splat())])that works.
This would be solved by the "clamp-to-last" data model that we've talked about in Stockholm (for which I still need to write an issue at some point...).
The following will not do what you'd expect, since it doesn't know that the main component has multiple instances:
Worse, there's no public way to explicitly indicate that we want it to splat.
The only way is to go through multiple layers of private APIs:
and even then, the above does not work, you need to recast it some more for some reason:
that works.
This would be solved by the "clamp-to-last" data model that we've talked about in Stockholm (for which I still need to write an issue at some point...).