Hello everybody,
I am facing an issue and I needed help. As you can see in the script, I have created a 2d grid of panels with a thickness using lunch box. then I have added a chamfer to all panel edges. Now I want to give different chamfer values for inside and out side edges. But I can’t find out how can I distinguish the indices of inner edges (by inner I mean edges in each box that is touching a neighboring box) from outside edges.
question.gh (8.8 KB)
It wasn’t 100% clear to me if you intend to get all the breps’ edges that exist within the regions you circled in your screenshot or only the edges that would form exterior edge loops; in reality, there aren’t any actual naked edges in these geometries.
Anyway, GH has some really neat components to work with brep edges (edges from points, linearity, direction, etc.).
I don’t have lunchbox instlaled, but since you were simply making subsurfaces here’s an example:
question.gh (26.0 KB)
NOTE:
In this example, all the breps’ edges within your circled regions are filleted.
Notify if that works.
There are multiple ways to go about it, however I used a quick ‘cheat’ by generating a mesh box with your same UV divisions.
Best
*{edit} NOTE 2:
For more clarity, it probably pays to edit your topic title. The first thing one gets from it is that you’re working with a grid, which you are, however that could imply just points, or lines. You’re working with surfaces that are getting extruded into solids with no interior edges.
Not everyone has Lunchbox so it would be nice if you internalized the extruded panels.
The first FilEdge component is red because it’s doing outside edges and the middle square has none, but it works for all others. The second FilEdge (green group) does inside edges.
I can see a flaw on the top outside edges… 
P.S. Replaced version ‘a’ to add purple group that fixes a flaw on the top edges:
(moved reference curve up)
P.P.S. Found another flaw when the extrusions are thicker so implemented a different fix, projecting edge midpoints to the original surface plane before measuring their distance.
Simple is not always easy! Took me awhile to figure out a data tree to do this with only one FilEdge.
Way better. 
White group shows a better way to create the grid of rectangles:
And the purple group creates a “grid” of triangles; not sure why some of them are chopped off?
