The merged results (after Sen and André analyzed the patches) are under the directories defects4j
and humanevaljava
.
To analyze the patches we disagree upon, run manual_patch_analysis_martin.py
To run the scripts, you should:
# Setup environment
cd ./src/manual_analysis_martin/
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
python manual_patch_analysis_martin.py
Notes:
- If you consider a patch to be semantically equivalent, we skip the remainder of the patches for the same bug (if there are still any disagreements) to save time.