How diegoa-rodriguezc/sistemas-recomendacion works
Project type and purpose
• This “sistemas-recomendacion” repository is a full-stack, containerized movie-recommendation system. It evolved through three “Taller” exercises (Taller1: basic Flask; Taller2: FastAPI; Taller3: React + FastAPI + Docker). The final deliverable (Taller3) is a web application that:
• Serves a React-based UI to end users
• Exposes a FastAPI backend with endpoints for movies, genres, user ratings, recommendations
• Loads a hybrid recommendation model and data from local files
• Uses Firebase for authentication
• Is orchestrated via Docker Compose
File structure (focus on Taller3)
• Top-level:
• Dockerfile (builds the backend image)
• docker-compose.yml (orchestrates frontend and backend containers)
• package.json / vite.config.js (defines the React app)
• index.html (static host for React entry)
• Backend (./Taller3/backend):
• main.py (FastAPI app entrypoint)
• routers/ (six router modules: movies, genres, recommendations, user_ratings)
• models/ (data processing logic & hybrid recommender classes)
• data/ (movie & rating CSV, precomputed pickles)
• requirements.txt
• Frontend (./Taller3/src):
• components/ (React UI building blocks)
• views/ (page-level components: Home, Login)
• services/ (firebase.service.js for auth, movielensService.service.js for API calls)
• routes/, styles/, data/ (local mocks)
• No external DB container; data is file-backed and in-memory.
README for additional insights
• README is sparse but the file-tree shows clear separation: front vs. back, service abstractions, prebuilt model artifacts.