Enable Tracing Requirements and Source Code in Visual Studio Code using Large Language Models
| Typ | Bachelorarbeit | |
|---|---|---|
| Aushang | BA LLM VS Code Erweiterung.pdf | |
| Betreuer | Wenden Sie sich bei Interesse oder Fragen bitte an: Dominik Fuchß (E-Mail: dominik.fuchss@kit.edu, Telefon: +49-721-608-47381), Kevin Feichtinger (E-Mail: kevin.feichtinger@kit.edu) |
Motivation:
One of the biggest challenges for software engineers is understanding and navigating unfamiliar codebases. Particularly, if the source code and/or its documentation lacks clear links between the software requirements and its implementation. Thus, the often manual task of understanding how different code segments relate to a given requirement is time consuming and error-prone. Large Language Models offer a promising solution to bridge the gap between requirements and their corresponding implementation. Hence, changing the way developers interact with unfamiliar codebases and improving the comprehensibility of source code. This saves time and development costs, facilitates troubleshooting and speeds up the training period for new software developers.
Tasks:
In this thesis, a Visual Studio Code extension should be developed, which allows developers to trace textually defined requirements to a given software repository using a specified Large Language Model. To evaluate the extension, existing requirements to source code tracing benchmarks should be used and compared to existing state-of-the-art approaches.
Note
This thesis can be supervised in English and/or German.