can not processing mutag dataset

[AllenZhaoYB]

🐛 Bug

To Reproduce

Steps to reproduce the behavior:

1.mutag_dataset = Entities(root='./data/MUTAG', name='MUTAG')

Expected behavior

The dataset was processed

Environment

• OS: Ubuntu 20.04
• Python version: 3.7
• PyTorch version: 1.71
• CUDA/cuDNN version:cu110
• GCC version:9.3.0

Additional context

ValueError: invalid literal for int() with base 10: rdflib.term.URIRef('http://dl-learner.org/carcinogenesis#d331') This error occurs at torch_geometric.datasets.Entities. line 128. nodes_dict = {np.unicode(key): val for key, val in nodes_dict.items()} example of the keys is: rdflib.term.URIRef('http://dl-learner.org/carcinogenesis#d331').

[AllenZhaoYB]

I found the solution. It is because the depression of np.unicode. Just change the np.unicode to np.compat.unicode, it will work.