How dat0309/do-an-nay-bi-quy-am works
Project type and purpose
• This is a full-stack plant identification system combining:
• A Python ML/REST backend (FastAPI + YOLO/ONNX models + training pipeline)
• Two React front-end apps (a user-facing “clientapp” and an admin “dashboard”)
• A Flutter mobile app for cross-platform scanning
• Containerization (Docker + docker-compose + Nginx as API gateway)
File structure
• Top-level modules:
• backend/ – Python services & ML models
– FASTAPI-AGRICULTURE-IDENTITY & BE-agriculture-identity: REST endpoints for identify/classify
– yolo-identity-agriculture: abstraction over ONNX/PT models & processing functions
– train/: scripts & data for generating .h5 classification models
• frontend/
– clientapp/: React SPA for end users to upload or take plant images and view results
– dashboard/: React SPA for administrators—plants data, metrics, usage cards, sidebar
• mobile/plant_identification_app/: Flutter app (controllers → services → views widgets)
• dockerfiles/ & docker-compose.yml: service definitions, Nginx reverse proxy, multi-container setup
• nginx/: nginx.conf for routing API calls and serving static assets