zhongyu1998 / pg-gnn Goto Github PK

In COLLAB dataset, the first line is number of all graphs, and the second line is a graph and its labels.
I want konw the what does the next line mean.
and in you code
for j in range(n):
g.add_node(j)
row = f.readline().strip().split()
tmp = int(row[1]) + 2
if tmp == len(row):
row = [int(w) for w in row]
attr = None
else:
row, attr = [int(w) for w in row[:tmp]], np.array([float(w) for w in row[tmp:]])
if not row[0] in feat_dict:
mapped = len(feat_dict)
feat_dict[row[0]] = mapped
node_tags.append(feat_dict[row[0]])
why would you do that tmp = int(row[1]) +2

1. After entering the code “python main_tu.py --dataset IMDBBINARY --batch_size 16 --hidden_dim 16 --num_layers 5 --final_dropout 0.0 --graph_pooling_type sum --degree_as_tag --fold_idx ” into the command line window of Pycharm, a warning is displayed saying that there is a syntax error in this code. I found that this code seems to be a collection of functions and file names, not instructions. When trying to run the main_tu.py file, I also encountered no functions that required the above constraints to be entered in order to run. So I don't quite understand what you meant by writing this sentence. Hope you can explain it to me.

2. Although main_tu.py can be run directly after downloading, I am very confused about the output of this code file. If possible, I would appreciate your tips so that I can better understand the inputs and outputs of the model.

3. According to the paper, the accuracy of this model is its accuracy in judging movie types by analysis the networks of actors and actresses. Theoretically, this model is based on the relationship between the nodes and edges of the graph, but because the data set of this model is too abstract, it is all numbers and vectors, so I have no idea how this model is established and which one I should refer to. If possible, I hope you can provide me with a description of the data and hints on the code.

4. I am considering using the PS-GNN model for practice. For example, social network analysis. But I haven't found any suitable data set yet. I would be sincerely grateful if you have any recommendations.