test data set

[jl63]

When the model use the test data set to test the model, which parts of the codes are for getting the feature of the test data to cluster them? If I want to test the same codes with another test data set, do I need to include the publication information in the pubs_raw.json and recreate the author_feature.txt related to it? I am a little bit confused by why the feature.txt and pubs_raw.json seem to did not contain any information about test data set but we can still use the test data set to test the codes by simply using the ids in the test data set. We at least have to know the coauthor, venue and title information of the test data to cluster them right?
Thank you so much!