Regarding the issue of handling graph structures

[zhangzone]

Hello, I have reviewed your code and found that when processing the OAG raw dataset, the content contained in the nodes is paper_node={'id ': paper_id,' title ': title,' type ':'paper', 'time': int (time)}. However, the data. py does not specify the node type, only
self.node_forward = defaultdict(lambda: {})
self.node_bacward = defaultdict(lambda: [])
self.node_feature = defaultdict(lambda: [])
I don't quite understand the usage of these three methods, I would like to use your HGT. My graph structure includes a node set node=defaultdict (lambda: {}), each node including node ID, type, etc. The edge set is a list, and each edge is edg=[source_node], target_node, time = None, relation_type = None], Can you directly use your HGT? I see that the graph structure you are processing is different from that in data.Py, and I don't know which type to convert it to