README.md

metadata

title: Cooking Tutor
emoji: 👨‍🍳
colorFrom: yellow
colorTo: red
sdk: docker
sdk_version: latest
license: apache-2.0
short_description: CookTut with WebSearch, Memory, Multilingual

Cooking Tutor Backend

An intelligent cooking assistant that provides personalized recipe guidance, cooking techniques, and culinary tips with multilingual support (English, Vietnamese, Chinese).

🚀 Features

• Smart Recipe Recommendations: Get personalized recipes based on available ingredients
• Multilingual Support: English, Vietnamese (Tiếng Việt), and Chinese (中文)
• Cooking Techniques: Step-by-step guidance for various cooking methods
• Ingredient Substitutions: Smart suggestions for ingredient alternatives
• Dietary Accommodations: Support for various dietary preferences and restrictions
• Web Search Integration: Real-time cooking information from trusted sources
• Memory System: Contextual conversation continuity

🛠️ Technical Stack

• Backend: FastAPI with Python 3.9+
• AI Models: Google Gemini Flash API
• Memory: FAISS + Sentence Transformers for semantic search
• Translation: HuggingFace Transformers (VietAI/envit5-translation, Helsinki-NLP/opus-mt-zh-en)
• Web Search: DuckDuckGo + specialized cooking engines
• Deployment: Docker container on HuggingFace Spaces

🏃‍♂️ Quick Start

The API is automatically deployed and running on HuggingFace Spaces. You can interact with it through the web interface or API endpoints.

API Endpoints

• GET / - Health check
• POST /chat - Main chat endpoint
• GET /health - System health status

Example Usage

import requests

# Chat with the cooking tutor
response = requests.post("https://your-space-url.hf.space/chat", json={
    "user_id": "user123",
    "query": "How do I make perfect pasta?",
    "lang": "EN"
})

print(response.json())

🌍 Multilingual Support

The cooking tutor supports three languages:

• English (EN): Full feature support
• Vietnamese (VI): Complete Vietnamese language support
• Chinese (ZH): Simplified Chinese support

🔧 Environment Variables

Required environment variables for deployment:

FlashAPI=your_gemini_api_key

📚 API Documentation

Chat Endpoint

POST /chat

Request body:

{
    "user_id": "string",
    "query": "string",
    "lang": "EN|VI|ZH",
    "search_mode": true,
    "video_mode": false,
    "servings": 4,
    "dietary": ["vegetarian"],
    "allergens": ["nuts"],
    "equipment": ["oven", "stovetop"],
    "time_limit_minutes": 30,
    "skill_level": "beginner|intermediate|advanced",
    "cuisine": "italian",
    "structured": false
}

Response:

{
    "response": "string",
    "videos": [
        {
            "title": "string",
            "url": "string",
            "thumbnail": "string",
            "source": "string"
        }
    ]
}

🏗️ Architecture

┌─────────────────┐    ┌──────────────────┐    ┌─────────────────┐
│   Frontend      │    │   FastAPI        │    │   AI Models     │
│   (Vercel)      │◄──►│   Backend        │◄──►│   (Gemini)      │
└─────────────────┘    └──────────────────┘    └─────────────────┘
                              │
                              ▼
                       ┌──────────────────┐
                       │   Search &       │
                       │   Memory         │
                       │   Systems        │
                       └──────────────────┘

🔍 Search & Memory

• Web Search: Real-time cooking information from multiple sources
• Memory Management: Short-term and long-term memory for conversation context
• Content Processing: Advanced content extraction and summarization
• Citation System: Proper source attribution with inline citations

🚀 Deployment

This space is configured for Docker deployment on HuggingFace Spaces:

• Port: 7860
• Base Image: Python 3.9
• Auto-deploy: Enabled on push to main branch

📄 License

MIT License - see LICENSE file for details.

🤝 Contributing

Contributions are welcome! Please feel free to submit issues and pull requests.

📞 Support

For support or questions, please open an issue in the repository.