Excessive inv_edge_idx error when building graphs for 2D graphene structures using Siesta interface

[solar121381]

Hello, Professor Zhong
I’m currently using the Siesta interface of HamGNN to generate graph data for a 2D graphene slab (with vacuum along the z direction).However, during the graph construction step (graph_data_gen_siesta.py / read_siesta.py), the program continuously prints

inv_edge_idx error

like this picture

here are my hsx and fdf files

here is graph_data_gen_siesta.py poscar2siesta.py

graph_data_gen_siesta.py
poscar2siesta.py

I’m not sure whether this issue is caused by my settings in the scripts or if the reverse-edge errors naturally occur for 2D graphene structures. I’m looking forward to your guidance on this issue.

[QuantumLab-ZY]

This appears to be a bug in Siesta/HonPAS, rather than in graph_data_gen_siesta.py. When performing Siesta/HonPAS calculations using only the Γ-point, some edges (i→j) may not have the corresponding reverse edges (j→i).

[solar121381]

This appears to be a bug in Siesta/HonPAS, rather than in graph_data_gen_siesta.py. When performing Siesta/HonPAS calculations using only the Γ-point, some edges (i→j) may not have the corresponding reverse edges (j→i).

Dear Professor Zhong
My issue has been solved , as what you say, it is kpts' problem , thank you very much ！