Find the nearest polygon to a layout corner

JSS

Hi everyone,

I have a question about DRC.
I'm trying to figure out how to get the nearest polygon to a layout corner.
As you can see in the picture, I'm looking to get these four polygons.

Thanks for your help.

sekigawa

Hi,

How about this implementation with the KLayout Python Module?
I'm not using the DRC function.

JSS

Hi @sekigawa,

Thanks for the reply.
Your solution is great. It does solve my question.
But I'm wondering if there's a way to solve this with the DRC function.
Thanks again for getting back to me.

Best regards,

JSS

Matthias

@sekigawa - thanks for your nice solution as usual

And @JSS, no, there is no plain DRC solution as DRC is made for finding things in a specific distance, not the closest thing.

But DRC can be mixed with Ruby code. Inside a DRC script, you can eventually define your own functions. Like this:

def find_closest_to_corners(bbox, layer)

  # prepare a set of four corners and closest object collectors
  corners = RBA::DSimplePolygon::new(bbox).each_point.to_a
  closest = corners.collect { |c| [ nil, nil ] }

  # for each (merged) polygon on the input layer
  layer.each do |poly|

    # for each corner of the bounding box
    corners.each_with_index do |c,corner_index|

      # look if any point of the polygon is closer to any
      # closest point we found so far for this corner
      poly.each_point_hull do |p|

        # euclidian distance
        d = c.distance(p)

        # collect the closest objects
        if !closest[corner_index][0] || d < closest[corner_index][0]
          closest[corner_index] = [ d, poly ]
        end

      end

    end

  end

  # returns the closest objects
  result = polygons
  closest.each do |d,poly|
    poly && result.insert(poly)
  end

  result

end

report("Discussion 2578")

l1 = input(1, 0)    # border
l2 = input(2, 0)    # objects

find_closest_to_corners(l1.bbox, l2).output("Objects closest to corners")

Which gives this:

Please note however, that this function is slow on large cases.

Matthias

JSS

Hi @Matthias,

Thank you for your reply, it is perfect for my needs!

Cheers,

JSS