hi, sorrry to bother. I recently want to replicate the experiment of table 1 in paper。the paper says the varied graph results comes from doing experiment on a circle with 40 nodes which has been randomly attached by 3 type shapes each has 8 instances. I wonder if we can make sure the role id of each node under this situation? and what's the specific mean of "randomly", is it means we randomly select target node(no repeat) and attach the shape on it or just randomly select the shape(no repeat) and attach it to uniformly interval target nodes?