How diego16022/proyecto-integrador works
Project type and purpose
• This is a full-stack mobile application (“Outfit Recommender”) built with Flutter on the frontend and FastAPI on the backend.
• Purpose: capture a user’s skin tone with an AI model (ResNet34), manage a virtual closet, record usage metrics, and recommend outfits.
• It uses Amazon S3 for image storage and Amazon RDS (MySQL) for persistent data.
File structure
• Top level:
• Modelo/: stand-alone scripts for model training/inference (main.py, modelo_skin.py, predictor.py, recorte.py).
• backend/: FastAPI service.
• frontend/: Flutter multi-platform app (iOS, Android, web, desktop).
• backend/app:
• routers/: HTTP endpoints grouped by resource.
• crud/: data-access layer using SQLAlchemy.
• schemas/: Pydantic models for request/response validation.
• modelos/: PyTorch model loader and inference code, plus the .pt file.
• utils/: AWS S3 upload helper.
• database.py: database engine and session.
• main.py: FastAPI app entrypoint.
• frontend/lib:
• screens/: UI pages/widgets.
• services/: HTTP clients for auth, garments, outfits, recommendations.
• Config files: .env, requirements.txt, pubspec.yaml, Docker/gcloud not present but AWS credentials required.
README for additional insights
[Flutter App] <--> [FastAPI Backend] <--> [MySQL RDS]
|
└--> [ResNet34 Model]
└--> [Amazon S3]
• Tech stack: Flutter, FastAPI, PyTorch, SQLAlchemy, Pydantic, AWS S3 & RDS.