MediScribe AI (doc2patient)

A demo application showcasing the capabilities of Large Language Models (LLMs) in the medical field. The system supports doctors by automatically extracting data from a patient visit and generating an easy-to-understand summary for the patient.

Prerequisites

To run the project, you need:

• Installed Python 3.9+
• API Keys:
  • Google Gemini API Key (for Gemini models)
  • Groq API Key (for Llama models)

Step-by-step Installation

1. Cloning the repository or downloading the code

Navigate to the project folder in the terminal:

cd doc2patient

2. Creating a virtual environment (optional but recommended)

For Windows:

python -m venv venv
venv\Scripts\activate

For macOS / Linux:

python3 -m venv venv
source venv/bin/activate

3. Installing required libraries

Run the following command in the main project directory (or ensure you install all of them):

pip install streamlit pandas plotly python-dotenv google-genai groq openai pypdf python-docx

(Alternatively, if you have a requirements.txt file, run pip install -r requirements.txt)

4. Configuring environment variables

Create a .env file in the main application folder (or inside the app/ folder) where the app.py file is located. Add your API keys to it according to the following scheme:

GOOGLE_API_KEY=your_google_gemini_api_key
GROQ_API_KEY=your_groq_api_key

Running the application

Make sure you are inside the app folder where the app.py file is located. Run the following command:

cd app
streamlit run app.py

A local address (e.g., http://localhost:8501) will appear in the terminal, which will automatically open in your browser.

Project structure

• app/app.py - Main entry file for the Streamlit view

• app/pages/ - Individual application views:

  • doctor_panel.py - Doctor's panel facilitating the loading and processing of medical notes
  • patient_panel.py - Patient's panel showing an accessible summary of the visit
  • laboratorium.py - Analytical module facilitating the comparison of models on a built-in dataset

• app/utils/ - Helper files:

  • ai_engine.py - Logic for communication with Llama and Gemini models
  • translations.py - File with multilingual constants to support PL/EN versions

  📸 Screenshots

Main Application

Doctor Panel Facilitates the loading and processing of medical notes.

Patient Portal Generates an easy-to-understand summary of the visit for the patient.

Laboratory Module (Model Comparison) Analytical module for comparing different LLMs.