From 5f50a5c60f38cca83b913dcbd56c2aba9b0a2c3b Mon Sep 17 00:00:00 2001 From: tylerreed Date: Wed, 24 Jan 2024 16:19:23 -0500 Subject: [PATCH] started a small beginner course for AutoGen, with some of the updates and to go over again the basics of what you can do, and how to connect to OPENAI and perform tasks with agents --- autogen_beginner_course/.env | 2 + autogen_beginner_course/OAI_CONFIG_LIST | 6 +++ autogen_beginner_course/assistant.py | 18 +++++++ autogen_beginner_course/chatbot.py | 19 +++++++ autogen_beginner_course/groupchat.py | 71 +++++++++++++++++++++++++ 5 files changed, 116 insertions(+) create mode 100644 autogen_beginner_course/.env create mode 100644 autogen_beginner_course/OAI_CONFIG_LIST create mode 100644 autogen_beginner_course/assistant.py create mode 100644 autogen_beginner_course/chatbot.py create mode 100644 autogen_beginner_course/groupchat.py diff --git a/autogen_beginner_course/.env b/autogen_beginner_course/.env new file mode 100644 index 0000000..287c814 --- /dev/null +++ b/autogen_beginner_course/.env @@ -0,0 +1,2 @@ +OPENAI_API_KEY=sk-1111 +model=gpt-4 \ No newline at end of file diff --git a/autogen_beginner_course/OAI_CONFIG_LIST b/autogen_beginner_course/OAI_CONFIG_LIST new file mode 100644 index 0000000..83e10a1 --- /dev/null +++ b/autogen_beginner_course/OAI_CONFIG_LIST @@ -0,0 +1,6 @@ +[ + { + "model": "gpt-4", + "api_key": "sk-1111" + } +] \ No newline at end of file diff --git a/autogen_beginner_course/assistant.py b/autogen_beginner_course/assistant.py new file mode 100644 index 0000000..1723ce5 --- /dev/null +++ b/autogen_beginner_course/assistant.py @@ -0,0 +1,18 @@ +import autogen + + +def main(): + # If you have created an OAI_CONFIG_LIST file in the current working directory, that file will be used. + config_list = autogen.config_list_from_json(env_or_file="OAI_CONFIG_LIST") + + # Create the agent that uses the LLM. + assistant = autogen.AssistantAgent("assistant", llm_config={"config_list": config_list}) + + # Create the agent that represents the user in the conversation. + user_proxy = autogen.UserProxyAgent("user_proxy", code_execution_config={"work_dir": "coding", "use_docker": False}) + + user_proxy.initiate_chat(assistant, message="Plot a chart of NVDA and TESLA stock price change YTD.") + + +if __name__ == "__main__": + main() diff --git a/autogen_beginner_course/chatbot.py b/autogen_beginner_course/chatbot.py new file mode 100644 index 0000000..80828b9 --- /dev/null +++ b/autogen_beginner_course/chatbot.py @@ -0,0 +1,19 @@ +import autogen + + +def main(): + # For example, if you have created a OAI_CONFIG_LIST file in the current working directory, that file will be used. + config_list = autogen.config_list_from_json(env_or_file="OAI_CONFIG_LIST") + + # Create the agent that uses the LLM. + assistant = autogen.ConversableAgent("agent", llm_config={"config_list": config_list}) + + # Create the agent that represents the user in the conversation. + user_proxy = autogen.UserProxyAgent("user", code_execution_config=False) + + # Let the assistant start the conversation. It will end when the user types exit. + assistant.initiate_chat(user_proxy, message="How can I help you today?") + + +if __name__ == "__main__": + main() diff --git a/autogen_beginner_course/groupchat.py b/autogen_beginner_course/groupchat.py new file mode 100644 index 0000000..a581441 --- /dev/null +++ b/autogen_beginner_course/groupchat.py @@ -0,0 +1,71 @@ +import autogen +import dotenv + +dotenv.load_dotenv() + +config_list = autogen.config_list_from_dotenv( + ".env", + {"gpt-4": "OPENAI_API_KEY"} +) + +gpt4_config = { + "cache_seed": 42, # change the cache_seed for different trials + "temperature": 0, + "config_list": config_list, + "timeout": 120, # in seconds +} +user_proxy = autogen.UserProxyAgent( + name="Admin", + system_message="A human admin. Interact with the planner to discuss the plan. Plan execution needs to be approved " + "by this admin.", + code_execution_config=False, +) +engineer = autogen.AssistantAgent( + name="Engineer", + llm_config=gpt4_config, + system_message="""Engineer. You follow an approved plan. You write python/shell code to solve tasks. Wrap the code in a code block that specifies the script type. The user can't modify your code. So do not suggest incomplete code which requires others to modify. Don't use a code block if it's not intended to be executed by the executor. +Don't include multiple code blocks in one response. Do not ask others to copy and paste the result. Check the execution result returned by the executor. +If the result indicates there is an error, fix the error and output the code again. Suggest the full code instead of partial code or code changes. If the error can't be fixed or if the task is not solved even after the code is executed successfully, analyze the problem, revisit your assumption, collect additional info you need, and think of a different approach to try. +""", +) +scientist = autogen.AssistantAgent( + name="Scientist", + llm_config=gpt4_config, + system_message="""Scientist. You follow an approved plan. You are able to categorize papers after seeing their + abstracts printed. You don't write code.""", +) +planner = autogen.AssistantAgent( + name="Planner", + system_message="""Planner. Suggest a plan. Revise the plan based on feedback from admin and critic, until admin approval. +The plan may involve an engineer who can write code and a scientist who doesn't write code. +Explain the plan first. Be clear which step is performed by an engineer, and which step is performed by a scientist. +""", + llm_config=gpt4_config, +) +executor = autogen.UserProxyAgent( + name="Executor", + system_message="Executor. Execute the code written by the engineer and report the result.", + human_input_mode="NEVER", + code_execution_config={ + "last_n_messages": 3, + "work_dir": "paper", + "use_docker": False, + }, +) +critic = autogen.AssistantAgent( + name="Critic", + system_message="Critic. Double check plan, claims, code from other agents and provide feedback. Check whether the " + "plan includes adding verifiable info such as source URL.", + llm_config=gpt4_config, +) +groupchat = autogen.GroupChat( + agents=[user_proxy, engineer, scientist, planner, executor, critic], messages=[], max_round=50 +) +manager = autogen.GroupChatManager(groupchat=groupchat, llm_config=gpt4_config) + +user_proxy.initiate_chat( + manager, + message=""" +find papers on LLM applications from arxiv in the last week, create a markdown table of different domains. +""", +)