Code2doc is a python experiment to generate documentation from code using ML
- Clone this repo
- Build the python environment from train/environment.yml https://docs.conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html#creating-an-environment-from-an-environment-yml-file
- Activate the code2doc environment
- Run train/code2doc.py
- The data used comes from Github's CodeSearchNet and its hosted in an S3 bucket, thanks to https://github.com/github/CodeSearchNet
- The data used will be downloaded automatically
- The only model currently available to train is a simple LSTM Seq2Seq model
MIT