refactor: Unify world-to-canvas, node-to-canvas point transforms

[gselzer]

There are a few instances of world -> canvas, node -> canvas methods popping up in intersection code:

• Example 1
• Example 2

Would probably be good to write some shared code for many of these functions