Dataset and source code for ICSE#196

The dataset used for this study is in dataset_codex folder.

The scripts for comment classification are in dataloader.py, model.py and prediction.py.

The evaluation metrics are in rouge.py and cbleu.py.

The code for evaluating codex on multi-intent comment classification is in test_codex.py, including all the selection and re-ranking strategies.

The samples and the scores of the user study are in user_study_codex.java.