Skip to content

Latest commit

 

History

History
251 lines (184 loc) · 8.68 KB

QUICKSTART.md

File metadata and controls

251 lines (184 loc) · 8.68 KB

NY State AI Hackathon - Quick Start Guide

Get any accelerator running in 5 minutes. All accelerators include mock mode for offline development - no Azure services required to get started.

🚀 Quick Start (Any Accelerator)

Python Accelerators (1-5)

# 1. Navigate to accelerator
cd [Accelerator-Directory]

# 2. Create virtual environment
python -m venv venv

# 3. Activate (Windows)
venv\Scripts\activate
# OR Mac/Linux: source venv/bin/activate

# 4. Install dependencies
pip install -r requirements.txt

# 5. Run demo (mock mode)
python demo.py

# 6. Run web interface
python -m src.main

.NET Accelerator (6)

# 1. Navigate to accelerator
cd DotNet-Virtual-Citizen-Assistant

# 2. Restore packages
dotnet restore

# 3. Run (mock mode)
dotnet run --project VirtualCitizenAgent

# Open http://localhost:5000

📁 The 6 Accelerators

# Accelerator Directory Tests Demo
1 Constituent Services Agent Constituent-Services-Agent/ 43 python demo.py
2 Document Eligibility Agent Document-Eligibility-Agent/ 86 python demo.py
3 Emergency Response Agent Emergency-Response-Agent/ 62 python demo.py
4 Policy Compliance Checker Policy-Compliance-Checker/ 14 python demo.py
5 Inter-Agency Knowledge Hub Inter-Agency-Knowledge-Hub/ 38 python demo.py
6 Virtual Citizen Assistant DotNet-Virtual-Citizen-Assistant/ 22 dotnet run

Total: 265 tests passing

🧪 Running Tests

Python Projects

cd [Accelerator-Directory]
python -m pytest tests/ -v

.NET Project

cd DotNet-Virtual-Citizen-Assistant
dotnet test

🔧 Configuring Azure OpenAI (Optional)

All accelerators work in mock mode by default. To connect real Azure OpenAI services, follow these steps.

Step 1: Get Your Azure OpenAI Credentials

  1. Go to Azure Portal

  2. Navigate to your Azure OpenAI resource

  3. Click Keys and Endpoint in the left menu

  4. Copy:

    • Endpoint: https://your-resource-name.openai.azure.com
    • Key 1 or Key 2: Your API key

  5. Click Model deployments → Manage Deployments in Azure AI Studio

  6. Note your Deployment name (e.g., gpt-4, gpt-4o, gpt-35-turbo)

Step 2: Configure Python Projects (Accelerators 1-5)

Create a .env file in the accelerator directory:

# Required for all Python accelerators
USE_MOCK_SERVICES=false
AZURE_OPENAI_ENDPOINT=https://your-resource-name.openai.azure.com
AZURE_OPENAI_KEY=your-api-key-here
AZURE_OPENAI_DEPLOYMENT_NAME=gpt-4o
AZURE_OPENAI_API_VERSION=2024-02-15-preview

# Additional keys by accelerator:

# Document Eligibility Agent (Accelerator 2)
AZURE_DOCUMENT_INTELLIGENCE_ENDPOINT=https://your-doc-intel.cognitiveservices.azure.com
AZURE_DOCUMENT_INTELLIGENCE_KEY=your-key

# Inter-Agency Knowledge Hub (Accelerator 5)
AZURE_AI_SEARCH_ENDPOINT=https://your-search.search.windows.net
AZURE_AI_SEARCH_KEY=your-key
AZURE_TENANT_ID=your-tenant-id
AZURE_CLIENT_ID=your-client-id
AZURE_CLIENT_SECRET=your-client-secret

Step 3: Configure .NET Project (Accelerator 6)

Option A: Edit appsettings.json

Edit VirtualCitizenAgent/appsettings.json:

{
  "SearchConfiguration": {
    "UseMockService": false,
    "Endpoint": "https://your-search.search.windows.net",
    "IndexName": "citizen-services",
    "ApiKey": "your-search-api-key"
  },
  "OpenAI": {
    "UseMockService": false,
    "Endpoint": "https://your-resource-name.openai.azure.com",
    "ApiKey": "your-openai-api-key",
    "DeploymentName": "gpt-4o"
  }
}

Option B: Use Environment Variables

# Windows PowerShell
$env:OpenAI__Endpoint = "https://your-resource-name.openai.azure.com"
$env:OpenAI__ApiKey = "your-api-key"
$env:OpenAI__DeploymentName = "gpt-4o"
$env:OpenAI__UseMockService = "false"

# Mac/Linux
export OpenAI__Endpoint="https://your-resource-name.openai.azure.com"
export OpenAI__ApiKey="your-api-key"
export OpenAI__DeploymentName="gpt-4o"
export OpenAI__UseMockService="false"

Supported Azure OpenAI Models

Model Deployment Name Best For
GPT-4o gpt-4o Best quality, multimodal
GPT-4 gpt-4 High quality reasoning
GPT-4 Turbo gpt-4-turbo Fast, large context
GPT-3.5 Turbo gpt-35-turbo Cost-effective

Note: Use the deployment name you created in Azure AI Studio, not the model name.

🎯 Accelerator Use Cases

1. Constituent Services Agent

What it does: AI chatbot answering citizen questions about NY State services Try: "How do I apply for SNAP benefits?" / "How do I renew my driver's license?"

2. Document Eligibility Agent

What it does: Processes eligibility documents (W-2s, pay stubs, utility bills) Try: Upload sample documents from sample_documents/

3. Emergency Response Agent

What it does: Multi-agent planning for emergency response coordination Try: Simulate a hurricane or winter storm scenario

4. Policy Compliance Checker

What it does: Analyzes policy documents for compliance violations Try: Upload a sample policy document for analysis

5. Inter-Agency Knowledge Hub

What it does: Cross-agency search with permission-aware results Try: Search across DMV, DOL, OTDA knowledge bases

6. Virtual Citizen Assistant (.NET)

What it does: RAG-powered chatbot for NYC government services Try: Ask about housing, transportation, or city services

📋 Key Azure Services by Accelerator

┌─────────────────────────────────────────────────────────────────┐
│                    Common Services (All)                        │
├─────────────────────────────────────────────────────────────────┤
│  Azure AI Foundry  │  Entra ID  │  Azure OpenAI  │  Key Vault   │
└─────────────────────────────────────────────────────────────────┘

┌──────────────────┬──────────────────┬──────────────────┐
│ Constituent Svc  │ Document Elig    │ Policy Checker   │
├──────────────────┼──────────────────┼──────────────────┤
│ Foundry IQ       │ Doc Intelligence │ Azure OpenAI     │
│ Azure Translator │ Microsoft Graph  │ Blob Storage     │
│ Azure AI Search  │ Blob Storage     │ Azure AI Search  │
└──────────────────┴──────────────────┴──────────────────┘

┌──────────────────┬──────────────────┬──────────────────┐
│ Emergency Resp   │ Knowledge Hub    │ Virtual Citizen  │
├──────────────────┼──────────────────┼──────────────────┤
│ Multi-Agent SK   │ Foundry IQ       │ Semantic Kernel  │
│ Weather APIs     │ Azure AI Search  │ Azure AI Search  │
│ Azure Maps       │ Microsoft Graph  │ Azure OpenAI     │
└──────────────────┴──────────────────┴──────────────────┘

🏆 Demo Tips

  1. Mock Mode First: All accelerators work without Azure - start there
  2. Run Tests: Verify everything works with pytest tests/ -v or dotnet test
  3. Use Sample Data: Each accelerator includes test data
  4. Show Citations: Government AI must show sources - this is built-in
  5. Human-in-the-Loop: Demonstrate escalation and approval workflows

📚 Additional Resources

NY State AI Hackathon - January 2026