Better Context Makes Better Code Language Models: A Case Study on Function Call Argument Completion

This is the experiment code for our AAAI 2023 paper "Better Context Makes Better Code Language Models: A Case Study on Function Call Argument Completion".

Data

Please refer to https://registry.opendata.aws/pyenvs-and-callargs/ for downloading PYENVS and CALLARGS.

Code

Please check the sub-folder for more details:

generate folder contains the codes for generating PYENVS and CALLARGS.

training folder contains the codes for the experiments on CALLARGS.

Citation

You are more than welcome to cite our paper:

@article{pei2023better,
  title={Better context makes better code language models: A case study on function call argument completion},
  author={Pei, Hengzhi and Zhao, Jinman and Lausen, Leonard and Zha, Sheng and Karypis, George},
  journal={arXiv preprint arXiv:2306.00381},
  year={2023}
}