Python library dependencies:

• torch>=1.0.1
• scipy>=0.14.0
• numpy -v: 1.19.1
• gensim -v: 3.8.1
• NLTK -v: 3.5
• sklearn -v: 1.19.0

Dataset: Wasi Uddin Ahmad, Saikat Chakraborty, Baishakhi Ray, and Kai-Wei Chang. A transformer-based approach for source code summarization. ACL 2020, Online, July 5-10, 2020, pages 4998–5007. Association for Computational Linguistics, 2020.

• tf_idf_vec.py

  python tf_idf_vec.py 

You can get the TF_IDF vetor of input file, please place this input file into the data folder, and run:

• K-means.py

  python K-means.py 

In this way, the data that needs to be data augmentation can be obtained.

• eda_uad_bert_augment.py

• eda_uda_bert.py

  python eda_uad_bert_augment.py --input=train.txt --output=augmented.txt --num_aug=20

You can specify your own with --output. You can also specify the number of generated augmented sentences per original sentence using --num_aug.

• split.py

  python split.py
  

You need to split augmented.txt into javadoc.original and code.original_subtoken.