ReadMe.md

pip install gitpython

cd bugPredictor

/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)" brew install python python3 --version

python3 -m venv venv source venv/bin/activate

pip install -r requirements.txt

python3 data/count_commits.py python3 data/generate_commit_data.py python3 data/generate_bug_metadata.py

python llm/train_bug_predictor.py

python llm/predict_bugs.py --model-path checkpoints/model_epoch_5.pt --commit-message "Your commit message" --code-changes "Your code changes"

use generated model for prediction

1. Interactive Mode (easiest for testing):

python llm/predict.py --interactive --model-path llm/checkpoints/model_epoch_5.pt

2. Using JSON File:

Create a JSON file with commit data

echo '{ "message": "Fix null pointer exception in UserService", "code_changes": "public void processUser(User user) {\n if (user != null) {\n user.process();\n }\n}" }' > commit.json

Make prediction

python llm/predict.py --json-input commit.json --model-path llm/checkpoints/model_epoch_5.pt

3. Direct JSON Input:

python llm/predict.py --commit-data '{"message": "Fix bug in login", "code_changes": "if (user.isAuthenticated()) {...}"}' --model-path llm/checkpoints/model_epoch_5.pt