Initial commit: Medical AI Agent with Gemini and OpenMed datasets

.env.example

@@ -0,0 +1,2 @@
+GEMINI_API_KEY=your_key_here
+HF_TOKEN=your_token_here

.gitignore

@@ -0,0 +1,20 @@
+# Python
+__pycache__/
+*.py[cod]
+*$py.class
+venv/
+.env
+
+# Data
+data/cache/
+*.csv
+*.json
+*.parquet
+
+# IDEs
+.vscode/
+.idea/
+
+# OS
+.DS_Store
+Thumbs.db

README.md

@@ -0,0 +1,53 @@
+# Medical AI Agent
+
+A powerful, conversational medical agent built using **Google Gemini 1.5 Flash** and medical datasets from **Hugging Face (OpenMed)**.
+
+## 🚀 Features
+- **Conversational Intelligence:** Powered by Gemini 1.5 Flash for high-quality medical reasoning.
+- **Data-Driven:** Grounded in real medical dialogues and datasets from the [OpenMed Collection](https://huggingface.co/collections/OpenMed/medical-datasets).
+- **Dynamic Updates:** Built-in tools to fetch and integrate new datasets regularly.
+- **Web Interface:** Easy-to-use UI built with Gradio.
+
+## 📋 Prerequisites
+- Python 3.10+
+- Google Gemini API Key (get it from [Google AI Studio](https://aistudio.google.com/))
+- Hugging Face Token (if accessing private/restricted datasets)
+
+## 🛠️ Installation
+
+1. **Clone the repository:**
+   ```bash
+   git clone https://github.com/shahnewazkabirrafi017-hub/medical-ai-agent.git
+   cd medical-ai-agent
+   ```
+
+2. **Install dependencies:**
+   ```bash
+   pip install -r requirements.txt
+   ```
+
+3. **Set up environment variables:**
+   Create a `.env` file in the root directory:
+   ```env
+   GEMINI_API_KEY=your_gemini_api_key_here
+   HF_TOKEN=your_huggingface_token_here
+   ```
+
+## 🖥️ Usage
+Run the application locally:
+```bash
+python app.py
+```
+
+## 📂 Project Structure
+- `app.py`: The Gradio web interface.
+- `agent/`: Core AI agent logic.
+- `data/`: Scripts for dataset management and caching.
+- `requirements.txt`: Python package requirements.
+- `.env.example`: Template for environment variables.
+
+## 🤝 Contributing
+Updates are welcome! Since we plan to regularly add more datasets, feel free to submit pull requests with new data processing scripts.
+
+## ⚠️ Disclaimer
+This AI agent is for **educational and research purposes only**. It is not a substitute for professional medical advice, diagnosis, or treatment. Always consult with a qualified healthcare provider for medical concerns.

agent/core.py

@@ -0,0 +1,32 @@
+import google.generativeai as genai
+import os
+from dotenv import load_dotenv
+
+load_dotenv()
+
+class MedicalAgent:
+    def __init__(self, api_key=None):
+        self.api_key = api_key or os.getenv("GEMINI_API_KEY")
+        if not self.api_key:
+            raise ValueError("GEMINI_API_KEY not found. Please set it in your environment or .env file.")
+        
+        genai.configure(api_key=self.api_key)
+        
+        # Using Gemini 1.5 Flash for speed and generous free tier
+        self.model = genai.GenerativeModel(
+            model_name="gemini-1.5-flash",
+            system_instruction=(
+                "You are an advanced Medical AI Assistant. Your goal is to provide accurate, "
+                "helpful, and empathetic medical information based on available datasets. "
+                "Always maintain a professional tone. "
+                "IMPORTANT: You are an AI, not a doctor. Always include a disclaimer that "
+                "this information is for educational purposes and the user should consult "
+                "a real healthcare professional."
+            )
+        )
+        self.chat = self.model.start_chat(history=[])
+
+    def get_response(self, user_input, context=""):
+        prompt = f"User Question: {user_input}\n\nRelevant Medical Context/Data: {context}"
+        response = self.chat.send_message(prompt)
+        return response.text

app.py

@@ -0,0 +1,62 @@
+import gradio as gr
+from agent.core import MedicalAgent
+from data.loader import fetch_med_dialog_sample, format_dialogue_context
+import os
+from dotenv import load_dotenv
+
+load_dotenv()
+
+# Initialize the agent
+# Note: User needs to provide GEMINI_API_KEY in .env
+try:
+    agent = MedicalAgent()
+    # Fetch some initial context to 'prime' the agent
+    initial_data = fetch_med_dialog_sample(3)
+    medical_context = format_dialogue_context(initial_data)
+except Exception as e:
+    print(f"Warning: Agent initialization failed (likely missing API key): {e}")
+    agent = None
+    medical_context = ""
+
+def medical_chat_interface(message, history):
+    if not agent:
+        return "Error: Agent not initialized. Please ensure GEMINI_API_KEY is set in your .env file."
+    
+    try:
+        response = agent.get_response(message, context=medical_context)
+        return response
+    except Exception as e:
+        return f"An error occurred: {str(e)}"
+
+# Custom Theme for a Professional Medical Look
+theme = gr.themes.Soft(
+    primary_hue="blue",
+    secondary_hue="slate",
+    neutral_hue="slate",
+    font=[gr.themes.GoogleFont("Inter"), "ui-sans-serif", "system-ui", "sans-serif"],
+).set(
+    body_background_fill="*neutral_50",
+    block_background_fill="white",
+    block_border_width="1px",
+)
+
+with gr.Blocks(theme=theme, title="Medical AI Agent") as demo:
+    gr.Markdown(
+        """
+        # 🏥 Medical AI Assistant
+        Developed by **[shahnewazkabirrafi017-hub](https://github.com/shahnewazkabirrafi017-hub)**
+        
+        This agent is grounded in the **OpenMed** dataset collection and powered by **Gemini 1.5 Flash**.
+        
+        *⚠️ Disclaimer: This is an AI tool for educational purposes. Consult a doctor for medical advice.*
+        """
+    )
+    
+    chatbot = gr.ChatInterface(
+        fn=medical_chat_interface,
+        examples=["What are common symptoms of iron deficiency?", "How can I improve my sleep hygiene?", "Explain what hypertension is in simple terms."],
+        type="messages"
+    )
+
+if __name__ == "__main__":
+    demo.launch()

data/loader.py

@@ -0,0 +1,35 @@
+from datasets import load_dataset
+import pandas as pd
+import os
+
+def fetch_med_dialog_sample(limit=5):
+    """
+    Fetches a small sample from the MedDialog dataset to use as context.
+    """
+    print(f"Fetching {limit} dialogues from MedDialog...")
+    try:
+        # Loading a small portion of the English MedDialog dataset
+        dataset = load_dataset("OpenMed/MedDialog", split="train", streaming=True)
+        samples = []
+        for i, entry in enumerate(dataset):
+            if i >= limit:
+                break
+            samples.append(entry)
+        
+        return samples
+    except Exception as e:
+        print(f"Error fetching dataset: {e}")
+        return []
+
+def format_dialogue_context(samples):
+    context = "Here are some examples of medical dialogues for reference:\n\n"
+    for sample in samples:
+        # Adjust based on actual dataset structure (checking common fields)
+        dialogue = sample.get('utterances', sample.get('description', ''))
+        context += f"- {dialogue}\n"
+    return context
+
+if __name__ == "__main__":
+    # Test fetch
+    data = fetch_med_dialog_sample(2)
+    print(format_dialogue_context(data))

requirements.txt

@@ -0,0 +1,6 @@
+google-generativeai
+datasets
+pandas
+gradio
+python-dotenv
+huggingface_hub