AI Data Lake Migration Frontend

React frontend for the local ai-agent and dummy_sap services.

This repo is intentionally separate from the backend repos:

AI agent/
  ai-agent/              core FastAPI tool and agent server
  dummy_sap/             mock SAP S/4 OData server
  migration-frontend/    React UI

Prerequisite

Install Node.js LTS first if node --version or npm --version is not recognized.

Run Commands

Open three PowerShell terminals.

1. Dummy SAP Server

Use the port that worked on your machine. Example with 8011:

cd "C:\Users\Sadhana.Deshmukh\Desktop\AI agent\dummy_sap"
.\.venv\Scripts\python.exe -m uvicorn main:app --host 127.0.0.1 --port 8011 --reload

Check:

http://127.0.0.1:8011/

2. Core AI Agent Server

Make sure ai-agent\.env points to the dummy SAP port:

SAP_NGROK_BASE_URL=http://127.0.0.1:8011
SAP_API_USERNAME=demo
SAP_API_PASSWORD=demo123

Then run the core server on port 8000 (this is the port configured for the unified LLM endpoint):

cd "C:\Users\Sadhana.Deshmukh\Desktop\AI agent\ai-agent"
.\.venv\Scripts\python.exe -m uvicorn app.main:app --host 127.0.0.1 --port 8000 --reload

Check:

http://127.0.0.1:8000/health
http://127.0.0.1:8000/docs

3. React Frontend

The .env file is already configured. If you need to change the backend API URL, edit:

cd "C:\Users\Sadhana.Deshmukh\Desktop\AI agent\migration-frontend"
notepad .env

Verify the core API port matches your backend:

VITE_CORE_API_URL=http://127.0.0.1:8000

Install and run:

npm install
npm run dev

Open:

http://127.0.0.1:5173

UI Flow

The frontend features an intelligent, conversational chatbot that guides you through the migration setup:

Enhanced Conversational Experience

1. Intelligent Onboarding: The chatbot greets you with a friendly welcome and explains the migration process step-by-step.

2. Contextual Help System:

   • When asked for credentials or connection details, the chatbot detects if you need help
   • Offers detailed guidance on finding SAP URLs, client IDs, Azure storage details, etc.
   • Provides multiple help levels: quick overview, detailed step-by-step, and advanced guidance

3. Three-Step Configuration:

   • Source: Choose your SAP system (S/4HANA, ECC, Oracle, or Custom)
   • Cloud: Select your cloud provider (Azure, AWS, or GCP)
   • Destination: Pick the landing service (Azure Blob, ADLS Gen2, Synapse, or Databricks)

4. Smart Assistance:

   • "I need help finding these details" - Get step-by-step guidance
   • "I have the details" - Skip help and proceed
   • "Use demo/default" - Quick setup for testing

5. Real-time Guidance:

   • Finding SAP Base URL: Multiple methods including SAP GUI, Basis team contacts, IT documentation
   • Azure Storage Setup: Portal walkthrough, access key retrieval, container creation
   • Connection Strings: Security best practices and credential management tips

6. SAP Agent Integration:

   • Calls the backend unified LLM endpoint at http://127.0.0.1:8000/v1/llm/run with SAP domain routing
   • Natural language request processing for every typed chat request and setup action
   • SAP table-list requests are routed to the LLM/backend agent instead of frontend-only metadata
   • Approval-gated operations for data migration
   • Real-time status updates in chat

Traditional Features

• Collapsible AWS Glue-style catalog rail showing scanned tables
• Table selection with confidence scores and row estimates
• Backend approval workflow with approve/reject actions
• Persistent chat history throughout the session

Key Improvements

• More human-like interactions: The chatbot uses friendly, encouraging language
• Proactive help: Offers assistance before users get stuck
• Educational content: Teaches users where to find credentials and how systems connect
• Flexible paths: Users can get help, use defaults, or provide custom details
• Security awareness: Includes tips on credential management and best practices

Backend Integration

The frontend proxies API calls through Vite:

/api/health       -> CORE_API/health
/api/v1/llm/run   -> CORE_API/v1/llm/run

Table discovery and approval requests are routed through the unified LLM endpoint in SAP mode and use the backend approval flow.