The current type definitions specify that applyToPoint() requires a Point and applyToPoints() requires a Point[]. This is more restrictive than the actual implementations, which also allow [number, number] and [number, number][], respectively.

One way to fix this is to include a union of types in the function signatures, similar to what was done for fromTriangles() here:

Perhaps a better way would be to change the definition of Point to include object and array forms, if the intention is to support both forms everywhere. I'm suggesting something like this:

  type Point = { x: number; y: number } | [number, number];

The change would go here:

If this change is made, the function signature for fromTriangles() can be simplified.