title	emoji	colorFrom	colorTo	sdk	sdk_version	app_file	pinned
King Arthur Baking AI Assistant	🍰	blue	red	streamlit	1.28.0	app.py	false

🍰 King Arthur Baking AI Assistant

A sophisticated AI-powered assistant for King Arthur Baking mixes using RAG (Retrieval-Augmented Generation) with LangGraph, OpenAI, and MongoDB Atlas.

🚀 Features

• Intelligent Web Scraping: Automatically scrapes the King Arthur Baking mixes category
• RAG System: Combines semantic search with OpenAI embeddings for accurate product recommendations
• LangGraph Agent: Advanced AI agent with multi-step reasoning and tool use
• MongoDB Atlas Integration: Vector search and document storage
• Streamlit Frontend: Beautiful, interactive web interface
• Real-time Analytics: Product insights and data visualization

🛠️ Technology Stack

• Frontend: Streamlit with custom CSS styling
• AI Framework: LangGraph + LangChain
• LLM: OpenAI GPT-4o
• Embeddings: OpenAI text-embedding-3-small
• Database: MongoDB Atlas
• Web Scraping: BeautifulSoup4 + Requests
• Visualization: Plotly
• Deployment: Hugging Face Spaces

📦 Installation

1. Clone the repository:

   git clone <repository-url>
   cd king-arthur-baking-ai

2. Install dependencies:

   pip install -r requirements.txt

3. Set up environment variables: Create a .env file in the root directory:

   OPENAI_API_KEY=your_openai_api_key_here
   MONGODB_URI=mongodb+srv://username:password@cluster.mongodb.net/
   MONGODB_DB_NAME=king_arthur_baking_db
   MONGODB_COLLECTION_NAME=mixes
   

🔧 Configuration

The application uses several configuration parameters that can be adjusted in config.py:

• API Keys: OpenAI API key for LLM and embeddings
• Database: MongoDB Atlas connection string and collection names
• Scraping: Delay between requests and retry limits
• Agent: Temperature, max tokens, and model parameters

🕷️ Data Scraping

The scraper automatically extracts comprehensive product information:

• Product names and descriptions
• Pricing information
• Ingredients and nutritional details
• Instructions and features
• Product images and URLs
• Availability status

Running the Scraper

python scraper.py

This will:

1. Scrape all mixes from the King Arthur Baking website
2. Save data to mixes_data.json (avoiding duplicates)
3. Display scraping statistics

🗄️ Database Setup

MongoDB Atlas Configuration

1. Create a MongoDB Atlas account
2. Create a new cluster
3. Get your connection string
4. Update the MONGODB_URI in your .env file

Loading Data

python database.py

This will:

1. Connect to MongoDB Atlas
2. Load data from the JSON file
3. Generate embeddings for all products
4. Create necessary indexes

🤖 AI Agent

The LangGraph agent features:

• Query Analysis: Understands user intent and extracts key information
• Multi-Modal Search: Combines semantic and text search
• Product Recommendations: Personalized suggestions based on preferences
• Product Comparison: Side-by-side analysis of different mixes
• Reasoning: Advanced analysis and insights about products

Agent Workflow

1. Analyze Query: Determine intent and extract keywords
2. Route Decision: Choose appropriate search strategy
3. Search/Recommend/Compare: Execute the chosen action
4. Reasoning: Analyze results and provide insights
5. Response Generation: Create helpful, detailed responses

🎨 Frontend Features

The Streamlit interface includes:

• Chat Interface: Natural language interaction with the AI
• Agent Graph: Visual representation of the AI workflow
• Product Cards: Rich product information display
• Analytics Dashboard: Data insights and visualizations
• Control Panel: Scraping and embedding management

📊 Analytics

The application provides:

• Price Distribution: Analysis of product pricing
• Feature Analysis: Most common product features
• Database Statistics: Product counts and embedding status
• Search Performance: Query analysis and results

🚀 Deployment

Hugging Face Spaces

1. Create a new Space on Hugging Face
2. Choose "Streamlit" as the SDK
3. Upload your code files
4. Set environment variables in the Space settings
5. The app will automatically deploy

Local Development

streamlit run app.py

📋 Usage Examples

Basic Product Search

"I'm looking for a chocolate cake mix"

Recommendations

"Can you recommend some easy baking mixes for beginners?"

Product Comparison

"Compare different pancake mixes"

Ingredient Information

"What ingredients are in your bread mixes?"

🔍 API Reference

Main Classes

KingArthurScraper

• scrape_all_mixes(): Scrape all products from the website
• save_to_json(): Save data to JSON file avoiding duplicates

MongoDBManager

• insert_products(): Insert products into database
• search_products(): Text-based search
• semantic_search(): Embedding-based search

EmbeddingService

• generate_embedding(): Create embeddings for text
• semantic_search(): Find similar products
• hybrid_search(): Combine semantic and text search

KingArthurBakingAgent

• chat(): Main chat interface
• get_graph_visualization(): Get agent workflow graph

🛡️ Error Handling

The application includes comprehensive error handling:

• Connection Errors: Automatic retry with exponential backoff
• API Rate Limits: Proper delays and batch processing
• Data Validation: Input sanitization and type checking
• Graceful Degradation: Fallback options when services are unavailable

📈 Performance Optimization

• Batch Processing: Embeddings generated in batches
• Caching: Frequent queries cached for faster response
• Indexing: MongoDB indexes for efficient search
• Rate Limiting: Respectful scraping with delays

🤝 Contributing

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Add tests if applicable
5. Submit a pull request

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

• King Arthur Baking for their excellent products and website
• OpenAI for the powerful GPT-4o and embedding models
• MongoDB Atlas for the vector database capabilities
• Streamlit for the beautiful web interface framework
• LangGraph for the advanced agent workflow capabilities

🐛 Troubleshooting

Common Issues

1. Connection Errors: Check your internet connection and API keys
2. Rate Limiting: Reduce scraping frequency or batch sizes
3. Memory Issues: Limit the number of products processed at once
4. Authentication: Verify your MongoDB Atlas and OpenAI credentials

Support

For issues and questions:

1. Check the logs for detailed error messages
2. Verify all environment variables are set correctly
3. Ensure all dependencies are installed
4. Check the MongoDB Atlas connection and permissions

📝 Changelog

Version 1.0.0

• Initial release
• Web scraping functionality
• MongoDB Atlas integration
• OpenAI embeddings and chat
• LangGraph agent workflow
• Streamlit frontend
• Hugging Face Spaces deployment ready