Implement Question Paper Generator (Topic or PDF Input)

Build a module that generates question papers based on either a provided topic or an uploaded PDF. The system should extract or retrieve relevant content, generate multiple types of questions (MCQ, short, long, numerical), and output results in a structured JSON format.

Key Tasks:

- Implement PDF text extraction (PyMuPDF / Tesseract)
- Support topic-based question generation via LLM
- Include question types and difficulty levels
- Format output as JSON with fields: type, marks, difficulty, question, options, answer
- Add endpoint /generate in backend for question paper generation

Goal:
Enable users to automatically generate structured question papers using AI from either topic keywords or uploaded documents.