This repository contains a microservice-based backend architecture built to support AI-driven features such as face recognition and image embedding, similar to the functionality of the Fotoyu application. The backend is designed to be scalable, efficient, and ready for AI integration.
- Handles image uploads and compression (CPU-intensive)
- Isolated for scalability and performance
- Prevents blocking of I/O-bound services like Photo Service
- Manages image metadata and access
- Designed as I/O-focused microservice
- Communicates with Upload Service via gRPC
- Manages user transactions and related operations
- Independent domain to support future financial or logging features
- Handles authentication, user profiles, and user data
- Acts as a foundational service for all identity-based interactions
-
Scalability
CPU-heavy services (e.g., image compression in Upload Service) are separated to allow independent scaling. -
Concurrency Optimization
IO-bound (Photo Service) and CPU-bound (Upload Service) workloads are decoupled for efficient resource utilization. -
Domain-Driven Design
Clear separation of concerns across Upload, Photo, Transaction, and User domains for better maintainability and extensibility.
- Language: Go (Golang)
- Web Framework: Fiber
- Communication: gRPC
- Database: PostgreSQL
- Cache/Session: Redis
- Service Discovery: Consul
Integration of AI-based Face Recognition and Image Embedding is now available!
This feature enhances the system with intelligent photo processing and identity recognition, designed to integrate seamlessly with existing Upload and Photo services.
You can access and deploy the dedicated AI service from the following repository:
👉 AI Face Recognition Service – ai-yourmoments
This document describes the flow of the image upload and processing system across microservices.
- Client
- Upload Service
- User Service
- Photo Service
- AI Service
- Minio (S3)
-
Client → Upload Service
The client sends a request to upload an image to the Upload Service via HTTP/REST. -
Upload Service → User Service
The Upload Service authenticates the user by calling the User Service via gRPC, using a JWT token. -
User Service → Upload Service
The User Service authorizes the token and grants access, allowing the Upload Service to continue to the upload controller. -
Upload Service ↔ Photo Service & Minio (S3)
The Upload Service updates the photo metadata (RAW) in the database and uploads the raw image to Minio (S3) for storage. -
Upload Service → Client
The Upload Service starts a compression task and triggers the AI Service call asynchronously using a background goroutine. A successful upload response is returned to the client immediately, without waiting for AI processing results. -
Upload Service → AI Service (Async)
The AI Service receives the task via gRPC, downloads the compressed image from Minio (S3), and performs the following processing pipeline:- Face recognition
- Embedding generation
- Similarity calculation
-
AI Service ↔ Photo Service & Minio (S3)
The AI Service sends the similarity results to the Photo Service via gRPC. The processed image (e.g., with bounding boxes) is then uploaded back to Minio (S3).
- HTTP/REST for external API communication
- gRPC for inter-service communication
- Go routines for asynchronous processing
- Minio (S3-compatible) for object storage
This project is actively being developed and intended for future expansion with:
- AI capabilities
- Observability and monitoring tools
- CI/CD support
- Dockerized deployment