- We parse the input and output scenes into objects.
- Because we detect cases where multiple input objects (2) map to a single output object, we form “groups” in the input whereby we treat the pair of pixels as a single object with two components.
We build a mapping from input objects to output objects.
e.g.
- Objects with a width of 1 stay the same.
- Objects with a width of 10 (groups of two pixels) become a line.
- We notice that these input objects are “groups” of non-contiguous objects from the input, so we add a “Group” definition to the output to specify what groups of non-contiguous input objects should be combined as a single object.
- The two objects should both be pixels, their image should be the same as each other’s, as should be their Y values. Together as a group, they should form an object with a width of 10, and span from X=1 to X2=10, etc.
- As of September 2022, my rule finding code is finding a further simplification, not requiring different rules based on the width of objects, because even single pixels can be expressed as having their shape set to line/rectangle, since a pixel is just a line/rectangle with a width of 1.
