LLM Solution - Water Meter Reader

Overview

LLM Solution is a project that uses a combination of AI technologies and modern web development frameworks to read and process water meter data from images. The application integrates with Google Gemini to utilize generative AI for image recognition, processes the data with NestJS, and stores it in a PostgreSQL database. The entire solution is containerized using Docker for seamless deployment and scalability.

Technologies Used

• NestJS: A progressive Node.js framework for building efficient and scalable server-side applications.
• PostgreSQL: An open-source relational database to store and manage the water meter readings.
• Google Gemini AI: AI-powered image recognition to extract water meter numbers from images.
• Docker: Used to containerize the application, ensuring consistent environments for development and production.
• Swagger: Provides API documentation and testing interface, accessible at http://localhost:3000/api.

Directory Structure

The project is divided into several modules, each handling different responsibilities of the application:

├── customer
│   ├── dto
│   ├── exceptions
│   ├── repository
│   ├── serialize
│   └── types
├── generative-ai
│   └── (logic for Google Gemini AI integration)
├── measure
│   ├── dto
│   ├── exceptions
│   ├── repository
│   ├── serialize
│   ├── types
│   └── validate
├── shared
│   └── dto
├── storage
│   └── (handles file management and storage)
└── docker-compose.yml

Module Descriptions

• customer: Manages customer data, including types, data transfer objects (DTOs), exceptions, and serialization logic.
• generative-ai: Handles the integration with Google Gemini AI to read water meter images.
• measure: Manages the processing of water meter readings, including validation, repository logic, and serialization.
• shared: Contains shared DTOs used across the application.
• storage: Responsible for handling file uploads and managing storage.

Prerequisites

• Node.js (v16+)
• PostgreSQL (v12+)
• Docker (v20+)

Getting Started

1. Clone the repository:

   git clone https://github.com/yourusername/llm-solution.git
   cd llm-solution

2. Install dependencies:

   npm install

3. Set up environment variables:

   Create a .env file in the root directory and configure your database connection and API keys.

   DATABASE_URL=postgres://myuser:mypassword@db:5432/mydatabase
   GOOGLE_GEMINI_API_KEY=your-google-gemini-api-key

4. Run Docker:

   Ensure Docker is running, then use docker-compose to start the services.

   docker-compose up --build

5. Apply Prisma Migrations:

   After starting the container, apply the migrations to the PostgreSQL database:

   docker-compose exec app npx prisma migrate deploy

6. Access the application:

   The application will be running at http://localhost:3000.

API Endpoints

The API exposes several endpoints for managing customers and processing water meter readings.

• POST /upload: Upload a water meter image for reading..
• PATH /confirm: Confirme de measure value.
• GET /:id/list: Retrieves a list of measures
• GET /file/:id: Retrieve a temporary file

Swagger Documentation

Swagger provides an interactive API documentation at:

• URL: http://localhost:3000/api

You can use Swagger to explore the available API endpoints and test them directly from your browser.

Testing

To run tests, use the following command:

npm run test

Contributing

If you want to contribute to this project, feel free to open a pull request. We welcome any improvements and bug fixes.

License

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