codemining-treelm contains code for language models that work on trees.
codemining.ast contains code to convert ASTs to language-agnostic TreeNodes
codemining.lm contains an implementation of PCFGs and TSGs as well as some idiom-related code.
The project depends on three internal (maven) modules:
a) codemining-utils b) codemining-core c) codemining-sequencelm
The rest of the dependencies are declared in the maven dependencies.
This repository contains the code related to the paper:
@inproceedings{allamanis2014mining,
title={Mining Idioms from Source Code},
author={Allamanis, Miltiadis and Sutton, Charles},
booktitle={Proceedings of the 22nd ACM SIGSOFT International Symposium on Foundations of Software Engineering},
pages={472--483},
year={2014},
organization={ACM}
}
To train a TSG for Java use the main class in codemining.lm.tsg.tui.java.SampleBlockedTSG with the arguments
/path/to/folder binaryvariables filterblock 1.0 50
to run the TSG training as in the "Mining Idioms from Source Code" paper. For other options please explore the code.