Elia: AI-Powered GIS Assistant

Elia stands for Enhanced Location Intelligence Assistant. It is a demonstration project designed to showcase how artificial intelligence, specifically Google's Gemini LLM, can be integrated with geospatial technologies like ArcGIS Server and Google Maps to enhance geospatial workflows.

Elia is more than just a tool—it's an assistant that simplifies complex geospatial data by combining the power of modern AI with intuitive user interaction.

---

What is Elia?

Elia is an application that bridges the gap between GIS (Geographic Information Systems) and AI. It aims to:

• Explore how Google's Gemini LLM can function as a geospatial assistant.
• Integrate ArcGIS Server to analyze and describe geospatial data layers.
• Provide users with a dynamic interface that visualizes geospatial data on Google Maps.

---

Key Features

1. Gemini LLM Integration:

   • Leverage Google's cutting-edge large language model for natural language interaction.
   • Utilize function calling to process geospatial queries dynamically.

2. ArcGIS Server Support:

   • Allow users to input an ArcGIS Server URL.
   • Query the server to list and describe available layers.

3. Google Maps Integration:

   • Embed Google Maps for geospatial visualization.
   • Display ArcGIS Server data layers interactively.

4. Modern Architecture:

   • Front-End: React for a dynamic user interface.
   • Back-End: FastAPI for a lightweight and efficient Python server.

---

Planned Architecture

Monorepo Structure

/elia /client # React front-end /server # FastAPI back-end

Back-End (FastAPI)

• Handles communication with:
  • Google's Gemini API for AI functionality.
  • ArcGIS REST API for querying geospatial data.
• Processes user inputs (e.g., ArcGIS Server URLs) and provides layer descriptions.

Front-End (React)

• Provides an intuitive chat-based interface for interacting with Elia.
• Integrates Google Maps for data visualization.

---

How Elia Works

1. Input: Users provide an ArcGIS Server URL.
2. Processing:
   • Back-end queries the ArcGIS Server for layer information.
   • Gemini LLM interprets and describes the layers in user-friendly language.
3. Output:
   • The front-end displays layer descriptions.
   • Visualize layers on Google Maps (if available).

---

Why Elia?

Elia demonstrates how AI and GIS can work together to:

• Simplify the interpretation of complex geospatial datasets.
• Provide actionable insights through conversational AI.
• Bridge traditional GIS workflows with cutting-edge AI technology.

This project is a stepping stone toward smarter geospatial tools, showcasing the potential for enhanced decision-making powered by AI.

---

Setting up your dev environment

• Clone the repo down locally
• setup your .env:

  ENV_STATE=global
  DB_NAME=elia-api-db
  DB_HOST=localhost
  DB_PORT=5432 
  DB_SSL=prefer 
  DB_USER=postgres
  DB_PASSWORD=pa55word
  
  FRONTEND_URL=http://127.0.0.1:8080
  JWT_SECRET=somekindofsupersecretsecret
  GOOGLE_API_KEY=thisisobiouslyyourownapikeymakesureitsgotgeminiapi
  BIGQUERY_JSON_KEY_B64=base64 encoded string of the google service account .json
  GOOGLE_LLM_MODEL=gemini-1.5-pro-001
  CHAT_HISTORY_LIMIT=100
  

• setup python virtual environment
  • python -m venv .venv
• install dependancies with
  • python -m pip install -r requirements.txt
• start the database via docker using the docker-compose
  • docker-compose up -d
• start the local dev environment with (note that the port is 8001 because docker will run on 8000)
  • uvicorn elia_api.main:app --reload --port 8001

---

License

This project is licensed under the MIT License. See LICENSE for more details.