Using GPT-3.5 Turbo, the classification script helps filter and prioritize requirements, while the traceability script establishes links between documents to track dependencies. It includes prompts that are carefully crafted using 5 different prompt patterns to communicate effectively with the model and provides generated results for each pattern.
- Create an Open AI key here https://platform.openai.com/account/api-keys
- Add your Open AI key to the openai_keyt.txt file
- Select which task you want to run (either task_classify.py or task_traces.txt)
- The generations will be stored in the repies_taskname folder
The data used in this repository and corresponding paper come from:
- Cleland-Huang, J., Mazrouee, S., Liguo, H., & Port, D.. (2007). NFR.
- Ferrari, A., Spagnolo, G., & Gnesi, S. (2017). PURE: A Dataset of Public Requirements Documents. In 2017 IEEE 25th International Requirements Engineering Conference (RE) (pp. 502-505).