PropertySphere is a modern Property Management System built with ASP.NET Core, RavenDB, and a lightweight React frontend. It demonstrates RavenDB’s advanced features — GenAI integration, AI Agents, Spatial Indexing, Dynamic Indexing, Data Subscriptions, and Time Series — to deliver an intelligent, real‑time experience for renters and property managers.

Talk to your Telegram bot (replace with your bot’s handle). Try prompts like:

• "What were my utility charges last month?"
• "What’s my balance right now?"
• "Can you charge my card for pending rent?"
• "I need maintenance to fix a leaky faucet." (you can attach a JPG)
• "Any open service requests still pending?"

The bot can show balances, accept payments, open and track service requests, and react to photo uploads.

• AI‑Powered Renter Assistant: Telegram bot that lets renters report issues, check balances, and pay rent with natural language.

• Intelligent Maintenance (GenAI): Upload a photo (JPG) and a GenAI ETL generates a concise, actionable description for maintenance.

  

• Billing: Friendly flow for reviewing and paying outstanding debt items.

  

• Real‑time Notifications: RavenDB Subscriptions notify renters on analysis results and workflow updates.

• Location‑Aware: Spatial indexes enable querying by property/service-request locations.

• Financial Tracking: Debt items, payments, and utility usage with allocations.

1. .NET 8 SDK
2. RavenDB 7.1+ (local or RavenDB Cloud)
   • Getting started: https://docs.ravendb.net/7.1/start/getting-started#installation--setup
3. OpenAI API Key (required for AI features)
4. Telegram Bot Token (required for chat interface)
   • Create a bot via @BotFather → /newbot

1. Open RavenDB Studio (usually http://localhost:8080).
2. Create a database named PropertySphere.

Use appsettings.Development.json (recommended for local dev):

{
   "RavenDb": {
      "Url": "http://localhost:8080",
      "Database": "PropertySphere"
   },
   "Telegram": {
      "BotToken": "YOUR_TELEGRAM_BOT_TOKEN_HERE"
   },
   "AI_API_KEY": "YOUR_OPENAI_API_KEY_HERE"
}

Alternatively, set environment variables (PowerShell):

$env:AI_API_KEY = "sk-proj-..."             # OpenAI key (enables AI Agent & GenAI)
$env:Telegram__BotToken = "123456:ABC-DEF"  # Telegram bot token (note: double underscore)

Notes

• If AI_API_KEY is not set, AI Agent and GenAI tasks are skipped.
• If Telegram:BotToken is missing, the app runs but the Telegram polling service stays disabled.

Or use .NET User Secrets (recommended for local dev):

# Run in the project folder (contains PropertySphere.csproj)
dotnet user-secrets set "AI_API_KEY" "sk-proj-..."
dotnet user-secrets set "Telegram:BotToken" "123456:ABC-DEF"

User Secrets are loaded automatically in Development (see Properties/launchSettings.json) because this project already defines a UserSecretsId in PropertySphere.csproj.

dotnet restore

1. Open Command Palette → “Tasks: Run Task”.
2. Choose Run PropertySphere (or Watch PropertySphere for hot reload).

dotnet run --project .\PropertySphere.csproj

# Hot reload
dotnet watch run --project .\PropertySphere.csproj

App URL: http://localhost:5000 (serves the React frontend from wwwroot/).

To explore the app with realistic data and link your Telegram account:

1. Get your Telegram Chat ID: chat with @userinfobot and copy the numeric ID.
2. Use the VS Code task Generate Demo Data (you’ll be prompted for the Chat ID), or run:

Invoke-RestMethod -Uri "http://localhost:5000/api/datageneration/generate-data?telegramChatId=YOUR_CHAT_ID" -Method POST

# curl alternative
curl -X POST "http://localhost:5000/api/datageneration/generate-data?telegramChatId=YOUR_CHAT_ID"

Once generated, start chatting with your bot. Try “Hello” or “I have a leak in my kitchen”.

We configure a RavenDB AI Agent to answer renter questions using RAG over your RavenDB data (leases, debts, payments, utilities): looks up relevant documents, formats helpful responses, and exposes tool actions like creating service requests or charging a stored card.

• Code: services/PropertyAgent.cs
• Flow: Telegram → RavenDB Agent → Queries/Actions → Reply + follow‑ups

When a renter uploads a JPG via Telegram, the image is stored as an attachment on a Photo document. A GenAI task processes it and writes a concise description back to the document.

• Code: services/PropertyDescriptionGenerator.cs
• Mechanism: GenAiConfiguration + prompt → OpenAI model → updates Description

Subscriptions react after the GenAI description is saved, notifying the renter via Telegram with context.

• Code: services/PhotoSubscription.cs
• Query: all Photos where Description != null

We index service requests with geospatial fields for property‑centric queries.

• Code: Indexes/ServiceRequests_ByStatusAndLocation.cs
• Example (C#):

Map = requests => from sr in requests
                  let property = LoadDocument<Property>(sr.PropertyId)
                  select new {
                        Location = CreateSpatialField(property.Latitude, property.Longitude)
                  };

Utility usage (Water, Power) for each unit is stored as RavenDB Time Series and queried for billing and graphs.

• Ingest: demo generator writes hourly usage
• Query: AI Agent can request time‑bounded aggregations for renter units

• Quickstart: QUICKSTART.md
• Overview: SUMMARY.md