LiteMyMap - Local Setup Guide

Prerequisites

• Docker installed on your system

Environment Configuration

Create a .env file at LiteMapIntellisense_web/lite_map_intellisense/.env with the following required API keys:

Required Environment Variables

• AZURE_MAPS_CLIENT_ID - Your Azure Maps Client ID for authentication
• AZURE_MAPS_PRIMARY_KEY - Your Azure Maps Primary Key for API access
• GOOGLE_MAPS_API_KEY - Your Google Maps API Key for map services

Example .env File

# Sample .env file

AZURE_MAPS_CLIENT_ID="your_azure_maps_client_id"
AZURE_MAPS_PRIMARY_KEY="your_azure_maps_primary_key"
GOOGLE_MAPS_API_KEY="your_google_maps_api_key"

Note: Users are responsible for checking the license and obtaining permission from the provider for any models or datasets used in this project.

Quick Start

1. Build the Docker image:

   docker compose build

2. Start the application:

   docker compose up

The application will be available on localhost on port 8000. i.e http://localhost:8000

Required Files & Directories

Model Files

Place model files in LiteMapIntellisense_ai_server/models/

• Supported formats: .pt and .keras
• Plot Detector Model: Included as part of this research (paper pending publication)
• Row Detector Model: Included as part of this research (paper pending publication)
• YLS Disease Detector Model: Included as part of this research (paper pending publication)
• Additional model details and training datasets will be made available upon publication of the associated research paper

Visual Prompt Classes

Located in LiteMapIntellisense_ai_server/visual_prompt_classes/

• Requires matching .txt file and annotated image pairs

Output Directories

Directory	Purpose
LiteMapIntellisense_web/lite_map_intellisense/results/uploads/	Uploaded GeoTIFF files
LiteMapIntellisense_ai_server/auto_generated_safe_to_delete/	Temporary auto-generated images (safe to delete)

Roadmap

• Docker image export
• Implement flexible workflows
• UI cleanup and feature improvements
• QGIS plugin testing and refinement
• User Documentation
• New Field Testing

Additional Information

Docker Compose Architecture

The docker-compose.yml file defines a multi-container application with the following services:

1. Redis Service (litemapintellisense_redis):

   • Builds from the LiteMapIntellisense_redis directory.
   • Exposes port 6379 for Redis.

2. Web Service (litemapintellisense_web):

   • Builds from the LiteMapIntellisense_web/lite_map_intellisense directory.
   • Exposes port 8000 for the web application.
   • Depends on Redis and the AI server for startup.
   • Uses environment variables for Django settings and Celery broker URL.

3. Celery Worker (litemapintellisense_celery):

   • Also builds from the same directory as the web service.
   • Runs a Celery worker for background tasks.
   • Depends on Redis and the AI server.

4. AI Server (litemapintellisense_aiserver):

   • Builds from the LiteMapIntellisense_ai_server directory.
   • Exposes port 5000 for AI-related services. (Use :5000 if you don't want to expose localhost:5000 )
   • Includes a health check to ensure the service is running.

5. Nginx Service (litemapintellisense_nginx):

   • Uses the latest Nginx image.
   • Exposes port 8000 and serves static files.
   • Depends on the web and AI server services.

This architecture allows for a modular setup where each component can be developed and scaled independently.

Troubleshooting

• If you encounter issues with Docker, ensure that it is running and properly configured.
• Check the logs for any error messages by running:

  docker compose logs

Contributing

We welcome contributions! Please fork the repository and submit a pull request with your changes.

License

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

Cite This Work

If you use this project in your research or work, please cite it as follows:

Janardhan Vignarajan. (2025). LiteMyMap:  Cloud-Based AI Inference Platform for GIS Applications. Retrieved from [https:github/janadh/lite-my-map]

The published paper will be available for citation here once it is released.

Note

Additional Notes

If you want visual prompt and text prompt, you can download the following models:

• YOLOE-11L Segmentation Model: YOLOE (for visual prompt/text prompt segmentation)
• YOLOv8L World Model: YOLO-World (from fixed categories)