Creating edge_index for mesh graph

[atale480]

I'm creating own dataset. I have adjacency list representation of graph but PyTorch_Gometric need Graph connectivity in COO format for edge_index. I'm having 3D Mesh in off file format and off according to off file

mesh has 'x' number of edges.

I converted mesh to adjacency list format. Now how can I convert adjacency list to edge_index ?
I did it with following piece of code.

import open3d as o3d
import numpy as np

mesh = o3d.io.read_triangle_mesh(filename)
np_traingle = np.asarray(mesh.triangles)

tri = {}
for i in range(np_traingle.shape[0]):
key = np_traingle[i][0]
tri.setdefault(key, set())
tri[key].add(np_traingle[i][1])
tri[key].add(np_traingle[i][2])

  key = np_traingle[i][1]
  tri.setdefault(key, set())
  tri[key].add(np_traingle[i][0])
  tri[key].add(np_traingle[i][2])
  
  key = np_traingle[i][2]
  tri.setdefault(key, set())
  tri[key].add(np_traingle[i][0])
  tri[key].add(np_traingle[i][1])

tri = dict(sorted(tri.items()))
row1 = []
row2 =[]

for key, value in tri.items():
tmp1 = [key for x in range(len(value))]
row1.extend(tmp1)
tmp2 = list(value)
row2.extend(sorted(tmp2))

edge_index = [row1,row2]

But after executing this code edge_index shape comes [2, 2*x] (x is number of edges of in Mesh mentioned in off file)
I'm not sure wether I'm going in correct direction or not ?
Please let me know wether I'm correct or not and if not let me know how to convert mesh connectivity to edge_index.
Thanks in advance !!!
Following link contain one off file from my dataset
MeshFile

[rusty1s]

This looks correct to me since PyG expects that both directions of an edge are present in edge_index. You can also take a look at the FaceToEdge transform for this.

[atale480]

Okay. Thanks for the help.