Production RAG Application

Based on Booking.com's GenAI Agent Architecture

Project Structure

production_rag/
├── app/
│   ├── __init__.py
│   ├── main.py                 # FastAPI application entry point
│   ├── config.py               # Configuration management
│   ├── models.py               # Pydantic models
│   ├── agent/
│   │   ├── __init__.py
│   │   ├── orchestrator.py     # LangGraph agent orchestration
│   │   ├── tools.py            # Tool definitions
│   │   └── prompts.py          # Prompt templates
│   ├── guardrails/
│   │   ├── __init__.py
│   │   ├── pii_redactor.py     # PII detection and redaction
│   │   ├── topic_classifier.py # Topic classification
│   │   └── injection_detector.py # Prompt injection detection
│   ├── retrieval/
│   │   ├── __init__.py
│   │   ├── embeddings.py       # Embedding service
│   │   ├── vector_store.py     # Weaviate integration
│   │   └── template_retriever.py # Template retrieval tool
│   ├── evaluation/
│   │   ├── __init__.py
│   │   ├── llm_judge.py        # LLM-as-a-Judge evaluation
│   │   └── metrics.py          # Evaluation metrics
│   └── observability/
│       ├── __init__.py
│       └── tracing.py          # Arize/observability integration
├── tests/
│   ├── __init__.py
│   ├── test_guardrails.py
│   ├── test_retrieval.py
│   └── test_agent.py
├── docker/
│   ├── Dockerfile
│   └── docker-compose.yml
├── requirements.txt
├── .env.example
└── README.md

Quick Start

1. Copy .env.example to .env and fill in your API keys
2. Run docker-compose up to start Weaviate and the application
3. Access the API at http://localhost:8000
4. View docs at http://localhost:8000/docs