Error when running refinement.js script

[YanW23]

Hello, there is the following problem when I tried to run the refinement.js script for building a refined semantic model of s01-cb in task_01:

After executing the refinement.js script, I waited for 10 hours with no result, also the content in traversed_nodes.txt keeps increasing. I'm confused whether there is a possibility of an infinite loop, or the multi-edge weighted graph(MEWG) in task_01 is so complicated that it does take a lot of time to output the result.

Could you give me some guidance?

[giuseppefutia]

Thank you for your message!

Are you using your own data? If it is the case, could you give a sample?

[YanW23]

Thank you for your reply. In fact, I am using the data privided by you. Specifically, I proceed with the refinement process with the following command for building the refined semantic model of s01-cb in task_01: node run/refinement.js data/taheriyan2016/task_01/semantic_types/updated/s01-cb_st.json data/pc/model_datasets/scores/pc/6000/score.json data/taheriyan2016/task_01/semantic_models/steiner/graph/s01-cb_steiner.json data/taheriyan2016/task_01/semantic_models/plausibles/graph/s01-cb_graph.json data/pc/semantic_models/s01-cb

• data/taheriyan2016/task_01/semantic_types/updated/s01-cb_st.json is the semantic type file for s01-cb in task_01.

• data/pc/model_datasets/scores/pc/6000/score.json is the score file generated during the training at epoch 6000 for data source s01-cb. score.zip

• data/taheriyan2016/task_01/semantic_models/steiner/graph/s01-cb_steiner.json is the beautified version of the initial semantic model file for s01-cb in task_01.

• data/taheriyan2016/task_01/semantic_models/plausibles/graph/s01-cb_graph.json is the beautified version of weighted graph file including all plausible semantic models for s01-cb in task_01.

However, after more than 10 hours, there is still no expected output and the program keeps running:

• data/pc/semantic_models/s01-cb_refined.graph
• data/pc/semantic_models/s01-cb_refined_graph.json

At the same time, the content of the traversed_nodes.txt file is still increasing, and the size is about 250GB. I simply visualized s01-cb_graph.json, the beautified version of weighted graph file including all plausible semantic models for s01-cb in task_01 and found that the graph contains a large number of edges. . Therefore, I'm guessing if it's because the multi-edge weighted graph is too complicated. Maybe it's because of other reasons that the correct result is not generated, so I hope you can give me some guidance.