Inter-Agency-Knowledge-Hub

Inter-Agency Knowledge Hub

Cross-agency document search system with permission-aware results for NY State government. Provides unified search across multiple agency knowledge bases with role-based access control.

Quick Start

1. Setup Environment

# Create virtual environment
python -m venv venv

# Activate (Windows)
venv\Scripts\activate

# Activate (Mac/Linux)
source venv/bin/activate

# Install dependencies
pip install -r requirements.txt

2. Run Demo (Mock Mode)

No Azure services required:

python demo.py

3. Run Tests

python -m pytest tests/ -v

Features

• Unified Search: Query across 5+ agency knowledge bases
• Permission-Aware Results: Role-based filtering via Entra ID
• Citation Tracking: Source attribution for LOADinG Act compliance
• Cross-Agency References: Link related policies across agencies
• Human-in-the-Loop: Escalation for complex queries
• 7-Year Audit Logs: Complete search and access history

Supported Agencies

Agency	Domain	Documents
DMV	Transportation	Licensing, registration
DOL	Labor	Employment, wages
OTDA	Social Services	Benefits, assistance
DOH	Health	Public health, regulations
OGS	General Services	Procurement, facilities

Project Structure

Inter-Agency-Knowledge-Hub/
├── src/
│   ├── api/             # Flask routes
│   ├── config/          # Settings
│   ├── models/          # Search models
│   ├── services/        # Search, auth services
│   └── plugins/         # Semantic Kernel plugins
├── assets/              # Sample documents
├── data/                # Index data
├── tests/               # Test suite (38 tests)
├── demo.py              # Interactive demo
└── requirements.txt

Search Architecture

┌─────────────────────────────────────────┐
│            Search Request               │
│    (Query + User Context + Filters)     │
└───────────────────┬─────────────────────┘
                    │
         ┌──────────▼──────────┐
         │   Permission Check   │
         │   (Entra ID Groups)  │
         └──────────┬──────────┘
                    │
    ┌───────────────┼───────────────┐
    ▼               ▼               ▼
┌───────┐     ┌───────┐       ┌───────┐
│  DMV  │     │  DOL  │  ...  │  DOH  │
│ Index │     │ Index │       │ Index │
└───────┘     └───────┘       └───────┘

Configuration

For Azure services, create a .env file based on .env.example:

USE_MOCK_SERVICES=false

# Azure OpenAI (Required)
AZURE_OPENAI_ENDPOINT=https://your-resource-name.openai.azure.com
AZURE_OPENAI_KEY=your-api-key
AZURE_OPENAI_DEPLOYMENT_NAME=gpt-4o
AZURE_OPENAI_API_VERSION=2024-02-15-preview

# Azure AI Search (Required)
AZURE_AI_SEARCH_ENDPOINT=https://your-search.search.windows.net
AZURE_AI_SEARCH_KEY=your-key

# Entra ID Authentication (Required for production)
AZURE_TENANT_ID=your-tenant-id
AZURE_CLIENT_ID=your-client-id
AZURE_CLIENT_SECRET=your-client-secret

Where to find these values:

1. Azure OpenAI: Azure Portal → Your OpenAI resource → Keys and Endpoint
2. AI Search: Azure Portal → Your Search resource → Keys
3. Entra ID: Azure Portal → App registrations → Your app → Overview
4. Deployment Name: Azure AI Studio → Deployments (e.g., gpt-4o)

Search Modes

Mode	Description	Best For
Semantic	Vector similarity search	Natural language queries
Keyword	Traditional text matching	Exact phrases
Hybrid	Combined semantic + keyword	General use

Compliance Features

• LOADinG Act: Citation tracking for all AI responses
• RAISE Act: Transparent AI assistance disclosure
• Audit Trail: Complete 7-year query and access logs
• Data Privacy: PII masking in search results

Tech Stack

• Azure AI Search: Vector and hybrid search
• Foundry IQ: Intelligent retrieval
• Entra ID: Authentication and authorization
• Semantic Kernel: AI orchestration
• Flask: Web API framework
• Pydantic: Data models

Hackathon Team

NY State AI Hackathon - January 2026