- We parse the input and output scenes into objects.
- In the output scenes, the multi-color objects are parsed into a form that specifies their colored components:
In cases where an output object contains a component which was an object in the input, we produce a mapping between the objects, and specify an AddComponents transform.
e.g.
- We notice that if we group the object mappings by the color of their input object, the transformation applied is consistent.
- For red, a particular AddComponents transform is needed.
- For blue, a particular AddComponents transform is needed.
- For any other color, the output object is the same as the input object.



