RAG-Based Medical Assistant

Overview

This project develops a Retrieval-Augmented Generation (RAG) AI solution to assist healthcare professionals by providing quick access to reliable medical knowledge. It addresses challenges like information overload, enabling faster decision-making, improving diagnostics, and standardizing care practices. The system leverages the Merck Manuals, a comprehensive medical reference, to answer common medical queries such as diagnostic assistance, drug information, treatment plans, specialty knowledge, and critical care protocols.

The project is implemented as a Jupyter notebook (medical1.ipynb) and uses various Python libraries to process the Merck Manuals PDF, create a knowledge base, and generate responses to medical queries.

Features

• Diagnostic Assistance: Retrieve symptoms and treatments for conditions (e.g., "What are the common symptoms and treatments for pulmonary embolism?").
• Drug Information: Provide trade names and details for medications (e.g., "Can you provide the trade names of medications used for treating hypertension?").
• Treatment Plans: Suggest first-line and alternative treatments (e.g., "What are the first-line options for managing rheumatoid arthritis?").
• Specialty Knowledge: Outline diagnostic steps for specific fields (e.g., "What are the diagnostic steps for suspected endocrine disorders?").
• Critical Care Protocols: Deliver protocols for urgent scenarios (e.g., "What is the protocol for managing sepsis in a critical care unit?").

Prerequisites

• Python 3.13 or higher
• Git for version control
• A virtual environment (recommended)

Installation

1. Clone the Repository:

   git clone https://github.com/your-username/RAGbased_medical_assistant.git
   cd RAGbased_medical_assistant

2. Set Up a Virtual Environment (optional but recommended):

   python3 -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install Dependencies: The project requires several Python libraries. Install them using the following commands (as listed in the notebook):

   pip install llama-cpp-python
   pip install numpy pandas
   pip install huggingface_hub tiktoken pymupdf
   pip install langchain langchain-community chromadb

   • For GPU support with llama-cpp-python, ensure you have CUDA installed and uncomment the GPU installation command in the notebook.
   • Note: You may need to restart the Jupyter kernel after installing some packages, as mentioned in the notebook outputs.

4. Download the Merck Manuals: The project uses the Merck Manuals (a 4,000+ page PDF) as the data source. Place the PDF file (medical_diagnosis_manual.pdf) in the project directory. (Note: The PDF is not included in the repository due to licensing restrictions.)

Usage

1. Launch Jupyter Notebook:

   jupyter notebook

   Open medical1.ipynb in your browser.

2. Run the Notebook:

   • Execute the cells in order to install dependencies, process the Merck Manuals PDF, and set up the RAG system.
   • The notebook includes code to load the PDF, extract text, and prepare the knowledge base using langchain and chromadb.

3. Query the System: Once the setup is complete, you can use the system to answer medical queries. Example queries:

   • "What are the common symptoms and treatments for pulmonary embolism?"
   • "What is the protocol for managing sepsis in a critical care unit?"

Project Structure

• medical1.ipynb: Main Jupyter notebook containing the RAG implementation.
• medical_diagnosis_manual.pdf: The Merck Manuals PDF (not included; must be provided by the user).
• chroma_sqlite3_medical_db: Directory for storing the Chroma vector database (created during execution).

Contributing

Contributions are welcome! Please follow these steps:

1. Fork the repository.
2. Create a new branch (git checkout -b feature/your-feature).
3. Make your changes and commit (git commit -m "Add your feature").
4. Push to the branch (git push origin feature/your-feature).
5. Open a Pull Request.

License

This project is licensed under the MIT License. See the LICENSE file for details (to be added).

Acknowledgments

• The Merck Manuals for providing comprehensive medical knowledge.
• Libraries like langchain, chromadb, and pymupdf for enabling this project.