RAG-Powered Healthcare Assistant

An AI-driven healthcare assistant that utilizes Retrieval-Augmented Generation (RAG) with multilingual support in Darija, Arabic, and English to offer personalized health management. The app helps users with prescription scanning, health queries, multilingual support, and pharmacy locator.

Background and Problem Statement

Managing healthcare information effectively remains a challenge in today's world. Common issues include:

• Fragmented medical records
• Difficulty in obtaining personalized medical advice
• Language barriers in healthcare communication
• Difficulty locating nearby healthcare providers

Impact and Proposed Solution

The RAG-Powered Healthcare Assistant aims to solve these challenges by:

• Centralizing personal health records for easy access
• Leveraging RAG to provide intelligent, context-aware responses to health queries
• Enabling Prescription OCR for easy scanning and processing of prescriptions
• Offering multilingual support (Darija, Arabic, English) for diverse users
• Incorporating a Pharmacy Locator with geolocation to help users find nearby pharmacies and healthcare providers

Project Outcomes and Deliverables

This project delivers:

• A fully functional AI-powered healthcare assistant
• A multilingual chatbot for health queries
• An OCR system for prescription scanning and processing
• An integrated Pharmacy Locator with real-time geolocation
• Secure storage and retrieval of personalized medical records

Technology Stack

• Frontend:
  • Streamlit - Main web interface
  • Chainlit - Chat interface
• Backend:
  • Python 3.9+
  • LangChain - For AI/LLM orchestration
  • ChromaDB - Vector database for document storage
  • Ollama - Local LLM integration
  • OpenCV - Image processing
  • PyPDF & PDFPlumber - PDF processing
  • Unstructured - Document parsing
  • ElevenLabs - Text-to-speech capabilities

Prerequisites

Before running the application, ensure you have:

1. Python 3.9 or higher installed
2. Git installed
3. Ollama installed (for local LLM support)
4. Sufficient disk space for dependencies and document storage
5. A modern web browser

Installation and Setup

1. Clone the repository:

   git clone https://github.com/cstar0521/Healthcare-Assistant.git
   cd Healthcare-Assistant

2. Create and activate a virtual environment (recommended):

   python -m venv venv
   source venv/bin/activate  # On Windows use: venv\Scripts\activate

3. Install the required dependencies:

   pip install -r requirements.txt

4. Install and start Ollama (if not already done):

   # Follow Ollama installation instructions at: https://ollama.ai/
   # Pull the required model:
   ollama pull llama3.2

Installation and Setup

1. OCR Interface (for prescription scanning):

   streamlit run ocr.py

2. Pharmacy Locator Interface:

   streamlit run pharmascysol.py

3. Chainlit Chat Interface (for health queries):

   chainlit run app_chainlit.py

4. Streamlit Documents Assitant :

   streamlit run app_streamlit.py

5. Main Interface:

   streamlit run main.py

The application will be accessible through your web browser at the provided local URL.

Future Enhancements

• Expanding language support to more regions and dialects.
• Improving OCR accuracy and expanding prescription types supported.
• Adding integration with healthcare provider APIs for real-time health data.