RAG OpenAI Assistants in AG2

This notebook shows an example of the GPTAssistantAgent with retrieval augmented generation. GPTAssistantAgent is an experimental AG2 agent class that leverages the OpenAI Assistant API for conversational capabilities, working with UserProxyAgent in AG2.

import logging
import os

from autogen import LLMConfig, UserProxyAgent
from autogen.agentchat.contrib.gpt_assistant_agent import GPTAssistantAgent

logger = logging.getLogger(__name__)
logger.setLevel(logging.WARNING)

assistant_id = os.environ.get("ASSISTANT_ID", None)

llm_config = LLMConfig.from_json(path="OAI_CONFIG_LIST")
assistant_config = {
    "assistant_id": assistant_id,
    "tools": [{"type": "retrieval"}],
    "file_ids": ["file-AcnBk5PCwAjJMCVO0zLSbzKP"],
    # add id of an existing file in your openai account
    # in this case I added the implementation of conversable_agent.py
}

gpt_assistant = GPTAssistantAgent(
    name="assistant",
    instructions="You are adapt at question answering",
    llm_config=llm_config,
    assistant_config=assistant_config,
)

user_proxy = UserProxyAgent(
    name="user_proxy",
    code_execution_config=False,
    is_termination_msg=lambda msg: "TERMINATE" in msg["content"],
    human_input_mode="ALWAYS",
)
user_proxy.initiate_chat(gpt_assistant, message="Please explain the code in this file!")

gpt_assistant.delete_assistant()

Edit this page