This repository is a code-reference package for the dissertation:
Design and Evaluation of a RAG-Enhanced Small Language Model Chatbot for ENT-domain Question Answering Under Resource Constrained Systems.
Included files:
- build_time_bootstrap.py
- run_time_bootstrap.py
- app.py
- environment specifications
- aggregate evaluation figures and tables
- build-time aggregate reports
Restricted files are not redistributed in this repository. These include:
- the source PDF
- the locally transcribed/reformatted PDF
- the ENT-QuAD dataset
- generated RAG lookup tables
- extracted raw images
- FAISS indexes
- embedding matrices
- row-level evaluation checkpoints
These materials are withheld because they contain or derive from third-party educational source material. The repository is intended to document the implementation logic and aggregate outputs, not to redistribute the restricted corpus-derived artefact.