Sidekick: Study Vibes, No Jive

Overview

Sidekick is an AI-powered study companion designed to enhance your learning experience and keep you motivated. It combines emotion recognition, task management, a personalized study guide generator, and AI-driven question answering to create a supportive and engaging study environment. The project integrates multiple advanced technologies including Streamlit for the user interface, DeepFace for real-time emotion detection, Hugging Face and Ollama for natural language processing, and MongoDB for data storage.

Features

• Emotion Detection: Uses your webcam and the DeepFace library to detect your current emotion (happy, sad, angry, neutral, fear, surprise, disgust). This information is used to provide tailored motivational messages.
• Task Management: Create and manage your study goals with a built-in to-do list. Mark tasks as complete and receive positive reinforcement.
• Personalized Study Guide Generator: Leverages a large language model (LLM) via Ollama to generate customized study guides based on your input. This includes key topics, suggested study approaches, resources, and time management tips.
• AI Study Assistant: Ask questions about your study material and get instant answers. This feature uses the Hugging Face mistralai/Mistral-7B-Instruct-v0.3 model for general-purpose question answering and provides enhanced capabilities through the OllamaLLM class for local model interactions.
• Course Assistant: Upload study materials (PDF, URL, or text) and ask specific questions about the content. This is powered by LangChain, Hugging Face embeddings, and Chroma vector database.
• Mood-Based Support: Get motivational messages tailored to your detected emotion to help you stay on track and overcome challenges.
• Progress Tracking: The app keeps track of your completed tasks and emotional states over time, allowing you to review your study patterns.
• Webcam Toggle: Ability to turn the camera on or off for privacy.

Technology Stack

• Streamlit: Web framework for creating the user interface.
• DeepFace: Facial analysis library for emotion detection.
• OpenCV: Used for webcam access and image processing.
• Hugging Face Transformers: mistralai/Mistral-7B-Instruct-v0.3 model for natural language processing tasks such as generating study guides and answering questions.
• Hugging Face Embeddings: sentence-transformers/all-MiniLM-L6-v2 for creating sentence embeddings for efficient text similarity searches.
• LangChain: Framework for building applications with large language models.
• Ollama: For local LLM interaction and enhanced language processing capabilities.
• Chroma: Vector database for storing and retrieving document embeddings.
• MongoDB: Database to store user data, including to-do lists and detected emotions.
• Python: The primary programming language.

Project Structure

• app.py: Main application file containing Streamlit UI code and logic for switching between pages.
• components/:
  • emotion_detector.py: Handles emotion detection using DeepFace and provides mood-based feedback.
  • todo.py: Manages the to-do list functionality.
  • qa_generator.py: Processes study materials, creates an index, and answers questions based on the content.
  • llm.py: Contains the OllamaLLM class for interacting with the Ollama API for local language model processing.
• database_mongodb.py: Handles interactions with the MongoDB database.
• logger.py: Provides logging functionality for debugging and monitoring.