AI Wardrobe Backend

AI-Wardrobe Frontend

https://github.com/giuliobarde/AI-Wardrobe-Frontend

Overview

The AI-Wardrobe backend is built with FastAPI and provides a RESTful API for managing virtual wardrobes, user authentication, outfit generation, and AI-powered clothing recommendations. The backend uses Supabase for data storage and authentication services.

Tech Stack

• Framework: FastAPI
• Database: Supabase
• Authentication: JWT-based auth via Supabase
• Image Storage: Supabase Storage
• LLM Integration: AI-powered outfit recommendations via Langchain

Key Features

• User authentication and profile management
• Virtual wardrobe management with detailed clothing attributes
• AI-powered outfit recommendations based on weather, occasion, and user preferences
• RESTful API with comprehensive endpoint coverage
• Secure data handling with proper authentication

Project Structure

backend/
├── fastapi/
│   ├── api/
│   │   ├── Database/       # Database interaction modules
│   │   ├── llm/            # LLM integration for outfit recommendations
│   │   ├── routers/        # API route definitions
│   │   ├── Weather/        # Weather integration services
│   │   ├── main.py         # FastAPI application entry point
│   │   ├── models.py       # Pydantic data models
│   │   └── test.py         # Test scripts

Documentation

For more detailed information about the API endpoints, database schema, and development guidelines, see the docs directory.

Setup and Installation

Prerequisites

• Python 3.8+
• Supabase account and project

Environment Variables

Required environment variables:

SUPABASE_URL=your_supabase_url
SUPABASE_ROLE_KEY=your_supabase_service_role_key
OPENAI_API_KEY=your_openai_api_key

Quick Start

1. Clone the repository
2. Create a virtual environment: python -m venv .venv
3. Activate the virtual environment
4. Install dependencies: pip install -r requirements.txt
5. Start the development server: uvicorn api.main:app --reload

API Documentation

When running the server, interactive API documentation is available at:

• Swagger UI: http://localhost:8000/docs
• ReDoc: http://localhost:8000/redoc

API Endpoints

Authentication

• POST /sign-up/: Register a new user
• POST /sign-in/: Authenticate a user and get session token
• GET /session/: Validate and get current session information
• POST /sign-out/: End the current user session

Clothing Items

• POST /add_clothing_item/: Add a new clothing item to user's wardrobe
• GET /clothing_items/: Get all clothing items for a user
• GET /clothing_item/{item_id}: Get details of a specific clothing item
• PUT /update_clothing_item/{item_id}: Update a clothing item
• DELETE /delete_clothing_item/{item_id}: Delete a clothing item

Outfits

• POST /create_outfit/: Create a new outfit from clothing items
• GET /outfits/: Get all outfits for a user
• GET /outfit/{outfit_id}: Get details of a specific outfit
• PUT /update_outfit/{outfit_id}: Update an outfit
• DELETE /delete_outfit/{outfit_id}: Delete an outfit

Profile

• GET /profile/: Get user profile information
• PUT /update_profile/: Update user profile information

AI Chat

• POST /chat/: Get AI-powered outfit recommendations

Database Schema

Users

Stores user authentication and profile information.

Clothing Items

Stores information about individual clothing items including:

• Type (top, bottom, outerwear, etc.)
• Material
• Color
• Formality
• Pattern
• Fit
• Weather suitability
• Occasion suitability
• Image reference

Outfits

Stores collections of clothing items that form complete outfits.

User Preferences

Stores user style preferences for AI recommendations.

Authentication Flow

The backend uses Supabase for authentication:

1. User signs up/signs in through the API
2. Supabase validates credentials and returns JWT
3. JWT is used for subsequent authenticated requests
4. Token validation happens through middleware

Development Guidelines

Adding New Endpoints

1. Create appropriate Pydantic models in models.py
2. Add database functions in the relevant module under Database/
3. Create route handlers in the appropriate router file
4. Register the router in main.py if creating a new router

Error Handling

All endpoints should use proper exception handling with appropriate HTTP status codes.

Testing

Run tests using the test scripts in test.py.