Closed cmavro closed 1 year ago
Thanks for your attention.
We follow the instruction from https://github.com/RichardHGL/WSDM2021_NSM/tree/main/preprocessing/Freebase.
Then we convert the entity name (relalation) from string to integer id.
I was referring to the subgraphs after path extraction as in this example: https://github.com/RUCKBReasoning/SubgraphRetrievalKBQA/issues/4#issuecomment-1235215817 I was able to find these subgraphs for Webqsp with the uploaded link in https://github.com/RUCKBReasoning/SubgraphRetrievalKBQA/issues/4#issuecomment-1236075450, but not for CWQ. Have you uploaded these subgraphs for CWQ by any chance?
Inside which, you can get the cache in npy format from this link https://drive.google.com/drive/folders/1ZHSBxGE-vjcvrfkmJPgcXfme_b2Ni6k6?usp=share_link.
Actually, I am looking for train_simple.json
(and so on) files for CWQ that contain the questions as well as the question-specific subgraphs. I found these for Webqsp in that link https://github.com/RUCKBReasoning/SubgraphRetrievalKBQA/issues/4#issuecomment-1236075450 in tmp/reader_data/webqsp
folder, but these are not present in the CWQ uploaded folder.
How can I find them? Btw, I am not looking for subgraphs extracted with PPR as in NSM, but the subgraphs extracted with your SR method. We simply want to test our KGQA model with your subgraphs. Thanks!
I upload a new file named "cwq_retrieved_graph_cache.zip" in that link.
Please note that the entity is converted to an integer~
Hope it can help you
Great, this is what I'm looking for, thanks a lot!
Hi, very interesting work.
How do I obtain the subgraphs for CWQ, i.e., the
tmp/reader_data/cwq
folder? It looks like this is not part of the zipped file uploaded in google drive.