AI-powered wardrobe and travel fashion companion. Recommends outfits based on your calendar, weather, wardrobe, and destination — whether that's the office tomorrow or a multi-city trip next month.

The core of the app is a multi-signal recommendation pipeline that combines:

• ML compatibility model — trained co-occurrence matrix scores category pairings
• Live weather — Open-Meteo forecast with automatic climatology fallback for distant-future trips or unavailable regions (historical averages from past 5 years)
• Cultural AI — Mistral-powered etiquette rules, local events, dress codes, and highlights for any destination
• Wardrobe matching — items the user owns are matched first; gaps produce shopping suggestions with links

Results are cached per signal (cultural 24h, weather 30min, shopping 6h) and I/O-bound fetches run in parallel via ThreadPoolExecutor.

• Add clothing items with photo, category, color, formality, season tags
• AI background removal for garment photos
• Track wear count and last-worn date
• Shared wardrobes — create collaborative closets with other users, add/remove members and items

• Calendar-driven outfit suggestions based on synced events
• Smart event classification — parses event titles/descriptions to infer formality (e.g. "board review" → polished, "spin class" → activewear)
• Morning outfit notification support
• Wear-again intelligence to avoid repetition

• Create trips with structured country + multi-city fields
• Per-city outfit recommendations run in parallel and display in tabbed UI
• Multi-day outfit plans with day-by-day weather, wardrobe matches, and gap analysis
• Packing checklist linked to wardrobe items
• Save/clear AI recommendations per trip
• Weather shows * with footnote when based on historical averages instead of live forecast

• Country + city destination search with place autocomplete (Open-Meteo geocoding, filtered by PCL/PPL feature codes)
• AI-powered cultural advice: dress code rules (severity-tagged), local events with clothing notes, must-visit places with what-to-wear tips
• Popular destination quick-select chips
• Tabbed result layout: etiquette, events, places, shopping

• Google Calendar, Outlook, and manual event sync
• Event type classification (meeting, workout, social, travel, wedding, interview, date, etc.)
• Formality auto-detection from event context

• User connections (follow/friend system)
• Shared wardrobes with member management
• Real-time messaging via WebSocket (Django Channels)

• CO₂ weight savings tracker
• Wear-again rewards / gamification
• Eco-shop integration for sustainable brand suggestions

• JWT authentication (access + refresh tokens)
• Email verification, password reset flow
• Google Calendar / Outlook / Apple Calendar connect/disconnect
• Push notification configuration

arokah/
├── backend/
│   ├── agents/             # AI agent views, services, Celery tasks
│   │   └── services.py     # daily_look, packing_list, conflict_detector,
│   │                        # cultural_advisor, outfit_planner, smart_recommend
│   ├── arokah/             # Django project settings, URLs, WSGI
│   │   └── services/
│   │       ├── recommendation_engine.py   # Unified rec engine (ML + weather + cultural)
│   │       ├── weather.py                 # Open-Meteo forecast + climatology fallback
│   │       ├── mistral_client.py          # Mistral AI wrapper
│   │       └── event_classifier.py        # Calendar event type classifier
│   ├── auth_app/           # User model, JWT auth, email verification
│   ├── wardrobe/           # ClothingItem CRUD, image upload
│   ├── itinerary/          # CalendarEvent, Trip (country/cities/JSONField), PackingChecklist
│   ├── calendar_sync/      # Google, Outlook, CalDAV integrations
│   ├── outfits/            # Outfit history, daily look records
│   ├── cultural/           # Cultural etiquette data
│   ├── social/             # User connections
│   ├── shared_wardrobe/    # Collaborative wardrobes
│   ├── messaging/          # WebSocket chat (Django Channels)
│   ├── sustainability/     # CO₂ tracker, eco-metrics
│   └── ml/                 # Fashion ML model (train.py, inference.py, artifacts/)
├── frontend/               # React 18 + Vite web app
│   └── src/
│       ├── pages/          # Dashboard, Wardrobe, TripPlanner, Cultural, People, etc.
│       ├── components/     # PlaceAutocomplete, shared UI
│       └── api/            # API client (fetch wrapper with JWT refresh)
├── mobile_flutter/         # Flutter mobile app
│   └── lib/
│       ├── screens/        # All app screens (mirror of web pages)
│       ├── widgets/        # PlaceAutocompleteField, shared widgets
│       ├── api/            # HTTP + WebSocket client with token management
│       └── theme/          # App colors, typography
└── requirements.txt        # Python dependencies

• Python 3.10+
• Node.js 18+
• Flutter SDK 3.11+

cd backend
python -m venv venv
source venv/bin/activate
pip install -r ../requirements.txt

cp .env.example .env   # Add MISTRAL_API_KEY, SECRET_KEY, etc.

python manage.py migrate
python manage.py createsuperuser
python manage.py runserver

cd frontend
npm install
npm run dev            # Vite dev server at localhost:5173

cd mobile_flutter
flutter pub get
flutter run

SECRET_KEY=…
MISTRAL_API_KEY=…           # Required for AI agents
REDIS_URL=redis://…         # Optional — enables Redis cache (else LocMem)
GOOGLE_CLIENT_ID=…          # For Google Calendar OAuth
GOOGLE_CLIENT_SECRET=…
MSAL_CLIENT_ID=…            # For Outlook OAuth
MSAL_CLIENT_SECRET=…

• POST /api/auth/register/ — Create account
• POST /api/auth/login/ — JWT token pair
• POST /api/auth/token/refresh/ — Refresh access token
• POST /api/auth/verify-email/ — Email verification
• POST /api/auth/forgot-password/ — Password reset request

• GET/POST /api/wardrobe/items/ — List / add clothing items
• PATCH/DELETE /api/wardrobe/items/:id/ — Update / remove item

• GET/POST /api/itinerary/events/ — Calendar events
• GET/POST /api/itinerary/trips/ — Trips (supports country, cities JSON fields)
• POST /api/itinerary/trips/:id/save-recommendation/ — Save AI recommendation
• DELETE /api/itinerary/trips/:id/save-recommendation/ — Clear saved recommendation
• GET/POST /api/itinerary/checklist/ — Packing checklist items

• POST /api/agents/daily-look/ — Today's outfit based on calendar + weather
• POST /api/agents/packing-list/ — Trip packing list generator
• POST /api/agents/outfit-planner/ — Multi-day outfit plan (5-4-3-2-1 capsule)
• POST /api/agents/cultural-advisor/ — Cultural etiquette + dress code advice
• POST /api/agents/conflict-detector/ — Schedule / weather / outfit conflict check
• POST /api/agents/smart-recommend/ — Unified recommendation (ML + weather + cultural); supports cities: [...] for parallel multi-city recommendations

• GET/POST /api/social/connections/ — User connections
• GET/POST /api/shared-wardrobes/ — Shared wardrobe CRUD
• WebSocket /ws/chat/:room/ — Real-time messaging

• Push notifications (morning outfit, trip reminders) — Firebase Cloud Messaging integrated
• Receipt-to-closet import — forward shopping emails to auto-populate wardrobe
• Outfit history & acceptance feedback — track which suggestions users accept/skip to improve ML model
• Apple Calendar (CalDAV) sync completion — backend scaffolded, OAuth flow pending

• Luggage weight predictor — estimate bag weight from item materials
• Multi-context day solver — outfit transitions for days with mixed events (e.g. office → gym → dinner)
• Social vibe analysis — scrape location tags for local style context
• Wear-again rewards gamification — streak tracking, badges
• Offline mode for mobile — cache wardrobe and last recommendations locally

• CO₂ weight savings calculator — carbon impact of lighter luggage
• Eco-commerce integration — prioritize sustainable / rental brands in shopping suggestions
• AI background removal for garment photos — server-side processing pipeline
• Style learning — personalized model fine-tuning from user feedback over time
• Docker Compose production deployment config
• Comprehensive test suite — unit + integration tests for agents and recommendation engine

MIT