This repo consists of scripts that perform operations on the great dataset.
To add noise to buggy samples, run add_noise.py as follows:
python3 add_noise.py <dir_name_of_sample> <noise_level_percentage> <trial_no> <estimate_of_total_no._of_buggy_samples>
We used the following bash command to run the script on multiple datasets:
for i in {1..5}; do time python3 add_noise.py train 25 ${i} 1798742 && zip -r train_noisy25_${i}.zip train_noisy25_${i} && rm -frv train_noisy25_${i}; done