Groq-Powered harmonyAPI

created for HackGT 12

build the demo here: demo repo

Description:

A comprehensive AI-powered Healthcare Provider (HCP) engagement API that provides intelligent medical literature analysis, risk prediction, and population health insights powered by ultra-fast Groq AI models. Transform clinical decision-making with evidence-based recommendations in seconds.

Basic Usage - Literature Search

curl -X POST https://your-api-domain.com/literature/search \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "specialty": "Cardiology",
    "keywords": ["heart failure", "statins"],
    "patient_conditions": ["hypertension"],
    "max_results": 3,
    "enable_ai_analysis": true
  }'

Response

{
  "status": "success",
  "data": {
    "studies": [
      {
        "title": "Combination ACE Inhibitor and Beta-Blocker Therapy in Heart Failure",
        "journal": "Journal of the American College of Cardiology",
        "publication_date": "2024-03-15",
        "relevance_score": 0.92,
        "authors": ["Johnson M", "Smith K"]
      }
    ],
    "ai_analysis": {
      "summary": "The literature strongly supports combination therapy for heart failure patients with comorbid hypertension.",
      "key_findings": [
        "Combination therapy reduces mortality by 28%",
        "Early initiation improves outcomes"
      ],
      "confidence_score": 0.89
    }
  }
}

Advanced Usage - Direct AI Analysis

Analyze any clinical text using Groq AI:

curl -X POST https://your-api-domain.com/ai/analyze \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "text": "65-year-old male with heart failure (EF 35%), hypertension, diabetes",
    "analysis_type": "clinical_implications",
    "model": "llama-3.1-8b-instant"
  }'

Response

{
  "status": "success",
  "data": {
    "analysis": "This patient requires evidence-based guideline-directed medical therapy. Key priorities include ACE inhibitor and beta-blocker therapy initiation.",
    "analysis_type": "clinical_implications"
  },
  "metadata": {
    "model_used": "llama-3.1-8b-instant",
    "next_steps": [
      "Review current medications for contraindications",
      "Order baseline labs (BUN, creatinine, electrolytes)"
    ]
  }
}

Risk Prediction

Assess patient cardiovascular risk:

curl -X POST https://your-api-domain.com/analytics/predict-risk \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "patient_data": {
      "age": 65,
      "systolic_bp": 150,
      "glucose": 130,
      "cholesterol": 260,
      "bmi": 32,
      "smoking": 1
    }
  }'

Response

{
  "status": "success",
  "data": {
    "risk_score": 0.78,
    "risk_level": "high",
    "10_year_risk_percentage": 23.4,
    "risk_factors": [
      {
        "factor": "hypertension",
        "severity": "moderate",
        "contribution": 0.25
      }
    ],
    "recommendations": [
      "Initiate antihypertensive therapy",
      "Start moderate-intensity statin therapy"
    ]
  }
}

AI Model Support

You can specify different Groq models for various use cases by using the model parameter:

curl -X POST https://your-api-domain.com/ai/analyze \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "text": "Complex clinical case requiring detailed analysis",
    "model": "llama-3.1-70b-versatile"
  }'

Supported AI Models

Model	Best For	Speed	Use Cases
llama-3.1-8b-instant	General analysis	Ultra-fast	Literature summaries, basic risk assessment
llama-3.1-70b-versatile	Complex reasoning	Fast	Differential diagnosis, detailed treatment planning
mixtral-8x7b-32768	Multi-specialty	Balanced	Cross-specialty consultations
gemma2-9b-it	Structured analysis	Fast	Guideline adherence, protocol development

Response Formats

You can request compact responses for mobile applications by adding the format parameter:

curl -X POST https://your-api-domain.com/literature/search?format=compact \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"specialty": "Cardiology", "keywords": ["heart failure"]}'

Response (Compact)

{
  "status": "success",
  "data": {
    "study_count": 5,
    "key_findings": ["Combination therapy reduces mortality by 28%"],
    "confidence": 0.89
  }
}

Authentication

All endpoints require authentication. First, obtain a token:

curl -X POST https://your-api-domain.com/auth/login \
  -H "Content-Type: application/json" \
  -d '{
    "username": "demo_provider",
    "password": "demo123"
  }'

Response

{
  "access_token": "eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9...",
  "token_type": "bearer",
  "user": {
    "username": "demo_provider",
    "role": "provider",
    "specialty": "Cardiology"
  }
}

Documentation

Local Development

Prerequisites

• Python 3.8+
• Groq API key (get one at console.groq.com)

Installation

git clone https://github.com/aandrx/harmonyAPI.git
cd harmonyAPI
pip install -r requirements.txt

Environment Setup

cp .env.example .env
# Edit .env with your configuration:
# GROQ_API_KEY=your_groq_api_key_here

Basic Usage

python app.py

The API will be available at http://localhost:5000

Testing

# Run comprehensive test suite
python test_api.py

Key Features

• Lightning-Fast AI: Ultra-fast Groq AI inference (sub-second responses)
• Medical Literature Search: PubMed integration with AI-powered relevance analysis
• Risk Prediction: Cardiovascular and health risk assessment
• Population Analytics: Health trend analysis across patient populations
• Multiple AI Models: Choose from Llama 3.1, Mixtral, and Gemma models
• Real-time Features: WebSocket support for live notifications
• Enterprise Security: JWT authentication with role-based access

API Endpoints

Endpoint	Method	Description
/auth/login	POST	Authenticate and get access token
/literature/search	POST	Search medical literature with AI analysis
/ai/analyze	POST	Direct AI analysis of clinical text
/analytics/predict-risk	POST	Patient risk prediction
/analytics/population-trends	POST	Population health analysis
/ai/models	GET	Available AI models
/health	GET	API health check

Dedication && Mission

This API is dedicated to healthcare providers worldwide who work tirelessly to improve patient outcomes through evidence-based medicine.

Our mission is to democratize access to AI-powered clinical decision support, making advanced medical analysis available to healthcare providers regardless of their organization's size or resources.

If you find this API helpful in your clinical practice, please consider:

Contributing to open healthcare initiatives

Sharing feedback to improve clinical workflows

Supporting medical education and research

Every API call represents a potential improvement in patient care. We believe that by providing healthcare providers with instant access to AI-analyzed medical literature and clinical insights, we can collectively raise the standard of care and improve health outcomes globally.

License

This project is licensed under the MIT License - see the LICENSE file for details.

---

Built with care for the healthcare community