- There seems to be a
width <= height capping. For example (height=200, width=500) produces the same box dimensions as (height=200, width=200).
- Though the bounding box change, the points still seem to roam beyond it in a square containing the rectangular box. Bounding box should bound the nodes, not just produce a limited view. Again, this happens in rectangular boxes.
Some code to try to see capping:
import numpy as np
from oui.splatter import _splatter
_splatter([{'fv': fv.tolist()} for fv in np.random.rand(100, 3)],
dict(nodeSize=3, height=200, width=900))
_splatter([{'fv': fv.tolist()} for fv in np.random.rand(100, 3)],
dict(nodeSize=3, height=900, width=200)

Some code to try to see that bounding box affects view but now actually "bounding".
import numpy as np
from oui.splatter import _splatter
_splatter([{'fv': fv.tolist()} for fv in np.random.rand(100, 3)],
dict(nodeSize=2, height=200, width=200))
_splatter([{'fv': fv.tolist()} for fv in np.random.rand(100, 3)],
dict(nodeSize=2, height=200, width=100))
_splatter([{'fv': fv.tolist()} for fv in np.random.rand(100, 3)],
dict(nodeSize=2, height=200, width=50))

width <= heightcapping. For example(height=200, width=500)produces the same box dimensions as(height=200, width=200).Some code to try to see capping:
Some code to try to see that bounding box affects view but now actually "bounding".