OpenAI Adapter for RAG Control

A high-performance OpenAI adapter for RAG (Retrieval-Augmented Generation) systems, providing seamless integration with rag-control framework for LLM and embedding operations.

Features

• 🚀 High-Performance LLM Generation - Support for chat completions with streaming
• 🔍 Query Embeddings - Generate embeddings for text queries using OpenAI's embedding models
• 📊 Comprehensive Metadata - Track latency, token usage, and request IDs
• ⚙️ Flexible Configuration - Full OpenAI client configuration support
• 🛡️ Error Handling - Robust exception handling with custom adapter errors
• 📝 Type-Safe - Full type annotations for Python static analysis

Installation

Install the package from PyPI:

pip install openai-adapter

Or with uv:

uv pip install openai-adapter

For development installation from source:

git clone <repo-url>
cd openai-adapter
uv sync

Version Compatibility

• Current Stable Version: v0.1.0
• Compatible with: rag-control v0.1.3

Quick Start

Integration with rag-control

The OpenAI Adapter is designed to integrate seamlessly with the rag-control framework:

from rag_control import RAGControl
from openai_adapter import OpenAILLMAdapter, OpenAIQueryEmbeddingAdapter

# Initialize adapters
llm_adapter = OpenAILLMAdapter(api_key="sk-...")
embedding_adapter = OpenAIQueryEmbeddingAdapter(api_key="sk-...")

# Create RAG system with OpenAI adapters
rag = RAGControl(
  llm_adapter=llm_adapter,
  embedding_adapter=embedding_adapter
)

# Use for RAG operations
response = rag.query("What is machine learning?")
print(response.answer)

API Reference

OpenAILLMAdapter

__init__(api_key: str, model: str = "gpt-3.5-turbo", **kwargs)

Initialize the LLM adapter.

Parameters:

• api_key (str): OpenAI API key for authentication
• model (str): Language model to use (default: gpt-3.5-turbo)
• **kwargs: Additional OpenAI client configuration
  • organization (str): Organization ID
  • project (str): Project ID
  • base_url (str): Custom API endpoint
  • timeout (float): Request timeout in seconds

Raises:

• LLMAdapterError: If client initialization fails

OpenAIQueryEmbeddingAdapter

__init__(api_key: str, model: str = "text-embedding-3-small", **kwargs)

Initialize the embedding adapter.

Parameters:

• api_key (str): OpenAI API key
• model (str): Embedding model (default: text-embedding-3-small)
• **kwargs: Additional OpenAI client configuration

Raises:

• QueryEmbeddingAdapterError: If client initialization fails

Development & Contributing

For information on setting up a development environment, code style guidelines, and contribution guidelines, see:

• DEVELOPMENT.md - Development environment setup and code quality tools
• CONTRIBUTING.md - Contributing guidelines and pull request process

License

Licensed under the Apache License 2.0. See LICENSE file for details.

Support

For issues and questions:

• Open an issue on GitHub
• Check existing documentation
• Review error messages for guidance