codekali/neural-sc-descriptor

Source Code Summarization

Neural Souce Code Descriptor

This is a deep learning project to summarize java methods into one line summary using state-of-the-art T5-Transformers.

Dataset

The dataset for this task has been taken from the DeepCom GitHub repository.

Repository: EMSE-DeepCom

Dataset: Gdrive link(select RQ3)

Usage

Clone this repo using git cli by git clone https://www.github.com/codekali/neural-sc-descriptor.git

Extract the downloaded dataset and create a folder named data in the cloned repo. Create folders train, val and test inside the data folder.

Move the train.sbt and train.nl files from the dataset extracted folder to the train folder. Do the same for val and test files too.

Training

Open the folder in a jupyter notebook and run the train.ipynb file.

Contributing

Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.

Please make sure to update tests as appropriate.

License

MIT

neural-sc-descriptor's Issues

Google Drive Link to dataset is broken

How does this work using T5 compare with DeepCommenter (https://github.com/SpringHerald/DeepCommenter)