Enriching Source Code with Contextual Data for Code Completion Models: An Empirical Study

This is material for a work under review. If used, please cite accordingly.

This is a reproduction package for "Enriching Source Code with Contextual Data for Code Completion Models: An Empirical Study", where line and token completion performance of UniXcoder, CodeGPT, and InCoder is assessed over datasets containing different amounts of type annotations and comments.