How awais-asghar/ai-based-exam-paper-generator-using-rag works
This repository is a notebook-based RAG exam-paper generator. Users provide a typed PDF and configure question types, counts, subject, and difficulty; the pipeline extracts and chunks text, retrieves context, generates and validates questions, then exports a formatted paper with an answer key. A parallel TF-IDF baseline and ROUGE-L/readability evaluation are also documented. Runtime components are represented at the notebook path because the notebook source was not sampled; the README documents their pipeline relationships.
Main entry points: ai-based-exam-paper-generator-using-rag.ipynb Notebook pipeline · exam_paper_RAG_difficult.pdf RAG exam PDF · exam_paper_BASELINE_difficult.pdf Baseline exam PDF
Components 17
Document Preparation
- Notebook pipeline
src/ai-based-exam-paper-generator-using-rag.ipynb - PDF extraction
src/ai-based-exam-paper-generator-using-rag.ipynb - Text chunking
src/ai-based-exam-paper-generator-using-rag.ipynb
Context Retrieval
- Dense embeddings
src/ai-based-exam-paper-generator-using-rag.ipynb - FAISS index
src/ai-based-exam-paper-generator-using-rag.ipynb - TF-IDF baseline
src/ai-based-exam-paper-generator-using-rag.ipynb - Context retrieval
src/ai-based-exam-paper-generator-using-rag.ipynb
Paper Generation
- Question generation
src/ai-based-exam-paper-generator-using-rag.ipynb - Output validation
src/ai-based-exam-paper-generator-using-rag.ipynb - PDF export
src/ai-based-exam-paper-generator-using-rag.ipynb
Comparison and Evaluation
- Metric evaluation
src/ai-based-exam-paper-generator-using-rag.ipynb - ROUGE and readability
src/ai-based-exam-paper-generator-using-rag.ipynb
Other components
- Exam creator
- Source PDF
- Groq LLaMA
- RAG exam PDF
output/exam_paper_RAG_difficult.pdf - Baseline exam PDF
output/exam_paper_BASELINE_difficult.pdf
Connections 22
- Exam creator → Notebook pipeline: configures
- Exam creator → Source PDF: provides
- Notebook pipeline → PDF extraction: runs
- Source PDF → PDF extraction: supplies text
- PDF extraction → Text chunking: passes text
- Text chunking → Dense embeddings: encodes chunks
- Dense embeddings → FAISS index: indexes vectors
- Text chunking → TF-IDF baseline: fits lexical index
- Notebook pipeline → Context retrieval: requests context
- Context retrieval → FAISS index: searches
- Context retrieval → TF-IDF baseline: searches baseline
- Notebook pipeline → Question generation: configures questions
- Context retrieval → Question generation: provides context
- Question generation → Groq LLaMA: prompts
- Groq LLaMA → Question generation: returns questions
- Question generation → Output validation: checks output
- Output validation → PDF export: provides paper
- PDF export → RAG exam PDF: writes PDF
- Notebook pipeline → Metric evaluation: runs comparison
- Metric evaluation → RAG exam PDF: scores RAG output
- Metric evaluation → Baseline exam PDF: scores baseline
- Metric evaluation → ROUGE and readability: computes scores