MCP Propertium Server

RestAPI + MCP server for OSM and Propertium data.

Features

• OpenAI integration for text processing and PDF analysis
• SQLAlchemy + Pydantic for database models and validation
• PostgreSQL with PostGIS for spatial data storage
• Planet Polygons API for geographic data

Getting Started

Prerequisites

• Docker and Docker Compose
• Python 3.12+

Installation

1. Clone the repository
2. Run with Docker Compose:

docker-compose up -d

Development Setup

# Install dependencies
uv sync
source .venv/bin/activate


uv run fastapi dev --app propertium.main:app ./src/propertium/main.py

API Endpoints

Planet Polygons

• GET /planet-polygons/ - Get all planet polygons (paginated)
• GET /planet-polygons/{osm_id} - Get a specific planet polygon by OSM ID
• POST /planet-polygons/ - Create a new planet polygon

OpenAI Processing

• POST /openai/pdf-to-text - Convert PDF to text using OpenAI
• POST /openai/text - Process text with OpenAI

Database Schema

The application uses SQLAlchemy with PostgreSQL and PostGIS for spatial data. The main table is fct_planet_polygons which stores geographic polygon data with the following fields:

• osm_id - Primary key, OSM ID
• admin_level - Administrative level
• boundary - Boundary type
• name - Name of the polygon
• z_order - Z-order for rendering
• way_area - Area of the polygon
• way - Geometry data (PostGIS)
• tags - JSON map of tags
• country - Country code
• avr_rent_price - Average rent price
• avr_rent_price_per_m - Average rent price per square meter
• avr_sale_price - Average sale price
• avr_sale_price_per_m - Average sale price per square meter
• is_city_for_search - Flag for search functionality
• name_en - English name
• code - Code
• alpha2 - Alpha-2 code
• border_type - Border type

Example Requests

Create a Planet Polygon

curl -X POST \
  http://localhost:8000/planet-polygons/ \
  -H "Content-Type: application/json" \
  -d '{
    "osm_id": 123456789,
    "admin_level": "8",
    "boundary": "administrative",
    "name": "Example City",
    "z_order": 0,
    "way_area": 1234.56,
    "tags": {"place": "city", "population": "100000"},
    "country": "US"
  }'