This repository is an artifact from the research paper:
Exploring Scientific Debt: Harnessing AI for SATD Identification in Scientific Software.
The notebook SSW_SATD_Analyzer.ipynb is a script that leverages a fine-tuned version of BERT-large for identifying SATD source code comments in scientific software. To use simply replace the line "https://github.com/healpy/healpy.git" in the notebook with your selected github repository that you would like to analyze. The script will then output relevant information pertaining to SATD in the respository along with saving a CSV file with all identified SATD instances and file paths.
For any questions or inquires please reach out to ericmelin@u.boisestate.edu.