Victorian Study App

Project Overview
Victorian Study App offers a comprehensive study companion experience by combining the following core functionalities:

User Authentication:
Secure sign-up, login, and logout features manage user sessions and credentials.

AI Chatbot:
Interact with "Sir Learnsworth," an AI tutor that answers study-related questions using advanced language models (e.g., GPT-4 via OpenAI) and context retrieved from user-uploaded knowledge bases.

Quiz Generation:
Quizzes are generated in two ways:

Knowledge Base-Based Quizzes:
Users can upload PDFs (or DOCX) files that are processed and split into text chunks. A vector store is built for context retrieval, and GPT-4 generates multiple-choice questions.
Conversation-Based Quizzes:
Based on ongoing interactions with the chatbot, a quiz is generated to assess retention and understanding.
Document Ingestion and Summarization:
The app processes various document formats (PDFs, DOCX, images) to extract text using libraries such as PyPDF2, pdf2image, pytesseract, and python-docx.

Streamlit Interfaces:
Standalone tools using Streamlit provide quick interactive summary and quiz generation via T5-based and LLaMA-based models.

Database Persistence:
User details and conversation histories are stored in Azure Cosmos DB, ensuring persistent and scalable data management.

Additional standalone Streamlit applications:

t5basedquizzgeneration.py: Uses a Flan-T5 model for PDF summarization and quiz generation.
llammabasedquiz-summarygeneration.py: Leverages an Open LLaMA-based model for generating summaries and quizzes from PDFs.
Technologies and Models Used
Backend Framework:

Flask: Web server and REST API endpoints.
Flask-CORS: Enables Cross-Origin Resource Sharing.
Jinja2: HTML templating engine.
Front-end:

HTML5, CSS3, JavaScript: Building responsive and interactive interfaces.
Streamlit: Rapid development of AI-based interactive tools.
AI and NLP:

OpenAI API (GPT-4): Conversational tutoring, quiz generation, and content summarization.
LangChain: For document ingestion, vector indexing, and managing conversational retrieval chains.
Transformers (Hugging Face): T5-based and LLaMA-based models for text generation in Streamlit apps.
PyPDF2, pdf2image, pytesseract, python-docx: Document processing and text extraction.
Vector Stores and Embeddings:

FAISS: Creating vector stores for efficient text retrieval.
OpenAIEmbeddings: Converting text into embeddings for context-based queries.
Database and Persistence:

Azure Cosmos DB: Stores user information and conversation histories.
Cosmos Python SDK: Managed via db.py to initialize containers for "users" and "conversations".
Configuration and Environment Management:

python-dotenv: Loads environment variables from a .env file.
Custom Config Class: Located in config.py (and app/config/config.py) for application settings.
Security:

Werkzeug Security: For password hashing and verification.
Installation and Setup
Prerequisites
Python 3.8+

Azure Cosmos DB:
Set up an Azure Cosmos DB account and obtain the database URL and key.

Environment Variables:
Create a .env file in the project root with the following variables:

SECRET_KEY=your_secret_key
OPENAI_API_KEY=your_openai_api_key
COSMOS_DB_URL=your_cosmos_db_url
COSMOS_DB_KEY=your_cosmos_db_key
COSMOS_DB_DATABASE=learnsworth

Running the Application Flask Server Start the Server: In the project root, run: python main.py Access the Web Interface:

Login/Signup: Visit http://0.0.0.0:8000/login to create an account or log in. Homepage: Upon login, the homepage (/) will display navigation links for the chatbot and quiz features. Chatbot: Accessible at /chatbot/chat_ui_page for interacting with Sir Learnsworth. Quiz: Accessible at /quiz/quiz_ui for quiz configuration and display. Streamlit Tools To run the standalone AI tools for PDF summarization and quiz generation:

T5-Based Quiz Generation:
streamlit run t5basedquizzgeneration.py
LLaMA-Based Quiz & Summary Generation: streamlit run llammabasedquiz-summarygeneration.py