Nicolo-Salvafiorita
54ae86a88a
- Added subscriptionPlan and subscriptionExpiresAt fields to UserProfile model. - Enhanced profile routes to handle subscription state updates and expiration logic. - Introduced email and password validation during user registration. - Implemented link stripping for processed articles in the articles route. - Added save article functionality with appropriate UI feedback in the frontend. - Updated AccountSettings and FeedContent components to reflect subscription state and saved articles. - Improved error handling and local storage management for user preferences.
📰 Digital Twin News
Sistema di News Intelligence basato su LLM che raccoglie notizie tramite RSS, le analizza con AI e fornisce una dashboard personalizzata per aziende.
Team
| Membro | Ruolo |
|---|---|
| Cipolla Diego | Data Pipeline Engineer |
| Bocca Lorenzo | Frontend & Auth Engineer |
| Galluzzo Matteo | AI, Modelli & Logica |
| Salvafiorita Nicolò | Integrazione AI/UI |
Architettura
[RSS Feed] → [Data Pipeline] → [MongoDB + Redis]
↓
[AI Analysis (Claude API)]
↓
[Backend API (Express)]
↓
[Frontend Dashboard (React)]
Stack tecnologico
| Layer | Tecnologie |
|---|---|
| Data Pipeline | Node.js, RSS Parser, Redis, node-cron |
| Storage | MongoDB, Redis |
| AI & Analytics | Python, Claude API, NLTK |
| Backend | Node.js, Express, JWT |
| Frontend | React, Material UI, Chart.js, Recharts |
| Infrastruttura | Docker Compose |
Avvio rapido
# 1. Avvia i database
docker-compose up -d mongo redis
# 2. AI Analysis
cd ai-analysis
source venv/bin/activate
python src/scheduler.py
# 3. Backend API
cd backend
npm run dev
# 4. Data Pipeline
cd data-pipeline
npm start
# 5. Frontend
cd frontend
npm start
KPI del sistema
- Frequenza aggiornamento: ogni 15 minuti
- Batch AI: 10 articoli per ciclo
- Autenticazione: JWT 24h
- Fallback locale: attivo se API LLM non disponibile
Limitazioni documentate
- Sistema batch-based (non real-time puro)
- Free tier API LLM gestito con fallback locale
- Testing automatizzato fuori scope MVP