E-APR (Explainable Automated Program Repair)

This repository contains the data and scripts of the paper:

E-APR: Mapping the Effectiveness of Automated Program Repair. Aldeida Aleti and Matias Martinez. 
arXiv:2002.03968  (2020)

If you use the data and/or scripts, please cite the paper.

@techreport{eapr2020,
 title = {E-APR: Mapping the Effectiveness of Automated Program Repair},
 year = {2020},
 author = {Aldeida Aleti and Matias Martinez},
 url = {http://arxiv.org/pdf/2002.03968},
 number = {2002.03968},
 institution = {arXiv},
 }

Repository organization

Features Folder

We consider 3 sets of features: 1) object-oriented features from ckjm_ext, 2) java features from jct, and 3) code features from Coming.

Plots

We present 3 sets of plots:

Plot included in the paper

Footprints all repaired bugs from RepairThemAll: 5 bug benchmarks 11 repair tools

Additional plots from experiments not included in the paper

Footprints all correctly repaired bugs from RepairThemAll: 5 bug benchmarks 11 repair tools

Footprints all correctly repaired bugs from Defects4J bug benchmark and 16 repair tools

Contact:

Aldeida Aleti aldeida.aleti@monash.edu

Matias Martinez matias.sebastian.martinez@gmail.com