[ad_1]
The method begins with scaffolding the autonomous brokers utilizing Autogen, a software that simplifies the creation and orchestration of those digital personas. We will set up the autogen pypi package deal utilizing py
pip set up pyautogen
Format the output (optionally available)— That is to make sure phrase wrap for readability relying in your IDE comparable to when utilizing Google Collab to run your pocket book for this train.
from IPython.show import HTML, showdef set_css():
show(HTML('''
<model>
pre {
white-space: pre-wrap;
}
</model>
'''))
get_ipython().occasions.register('pre_run_cell', set_css)
Now we go forward and get the environment setup by importing the packages and establishing the Autogen configuration — together with our LLM (Giant Language Mannequin) and API keys. You need to use different native LLM’s utilizing companies that are backwards appropriate with OpenAI REST service — LocalAI is a service that may act as a gateway to your domestically working open-source LLMs.
I’ve examined this each on GPT3.5 gpt-3.5-turbo and GPT4 gpt-4-turbo-preview from OpenAI. You will have to contemplate deeper responses from GPT4 nonetheless longer question time.
import json
import os
import autogen
from autogen import GroupChat, Agent
from typing import Optionally available# Setup LLM mannequin and API keys
os.environ["OAI_CONFIG_LIST"] = json.dumps([
{
'model': 'gpt-3.5-turbo',
'api_key': '<<Put your Open-AI Key here>>',
}
])
# Setting configurations for autogen
config_list = autogen.config_list_from_json(
"OAI_CONFIG_LIST",
filter_dict={
"mannequin": {
"gpt-3.5-turbo"
}
}
)
We then must configure our LLM occasion — which we are going to tie to every of the brokers. This enables us if required to generate distinctive LLM configurations per agent, i.e. if we wished to make use of totally different fashions for various brokers.
# Outline the LLM configuration settings
llm_config = {
# Seed for constant output, used for testing. Take away in manufacturing.
# "seed": 42,
"cache_seed": None,
# Setting cache_seed = None guarantee's caching is disabled
"temperature": 0.5,
"config_list": config_list,
}
Defining our researcher — That is the persona that can facilitate the session on this simulated person analysis state of affairs. The system immediate used for that persona features a few key issues:
- Objective: Your function is to ask questions on merchandise and collect insights from particular person prospects like Emily.
- Grounding the simulation: Earlier than you begin the duty breakdown the record of panelists and the order you need them to talk, keep away from the panelists talking with one another and creating affirmation bias.
- Ending the simulation: As soon as the dialog is ended and the analysis is accomplished please finish your message with `TERMINATE` to finish the analysis session, that is generated from the
generate_noticeoperate which is used to align system prompts for varied brokers. Additionally, you will discover the researcher agent has theis_termination_msgset to honor the termination.
We additionally add the llm_config which is used to tie this again to the language mannequin configuration with the mannequin model, keys and hyper-parameters to make use of. We’ll use the identical config with all our brokers.
# Keep away from brokers thanking one another and ending up in a loop
# Helper agent for the system prompts
def generate_notice(function="researcher"):
# Base discover for everybody, add your personal extra prompts right here
base_notice = (
'nn'
)# Discover for non-personas (supervisor or researcher)
non_persona_notice = (
'Don't present appreciation in your responses, say solely what is important. '
'if "Thanks" or "You are welcome" are mentioned within the dialog, then say TERMINATE '
'to point the dialog is completed and that is your final message.'
)
# Customized discover for personas
persona_notice = (
' Act as {function} when responding to queries, offering suggestions, requested on your private opinion '
'or collaborating in discussions.'
)
# Verify if the function is "researcher"
if function.decrease() in ["manager", "researcher"]:
# Return the total termination discover for non-personas
return base_notice + non_persona_notice
else:
# Return the modified discover for personas
return base_notice + persona_notice.format(function=function)
# Researcher agent definition
title = "Researcher"
researcher = autogen.AssistantAgent(
title=title,
llm_config=llm_config,
system_message="""Researcher. You're a prime product reasearcher with a Phd in behavioural psychology and have labored within the analysis and insights trade for the final 20 years with prime inventive, media and enterprise consultancies. Your function is to ask questions on merchandise and collect insights from particular person prospects like Emily. Body inquiries to uncover buyer preferences, challenges, and suggestions. Earlier than you begin the duty breakdown the record of panelists and the order you need them to talk, keep away from the panelists talking with one another and creating comfirmation bias. If the session is terminating on the finish, please present a abstract of the outcomes of the reasearch examine in clear concise notes not firstly.""" + generate_notice(),
is_termination_msg=lambda x: True if "TERMINATE" in x.get("content material") else False,
)
Outline our people — to place into the analysis, borrowing from the earlier course of we will use the persona’s generated. I’ve manually adjusted the prompts for this text to take away references to the most important grocery store model that was used for this simulation.
I’ve additionally included a “Act as Emily when responding to queries, offering suggestions, or collaborating in discussions.” model immediate on the finish of every system immediate to make sure the artificial persona’s keep on process which is being generated from the generate_notice operate.
# Emily - Buyer Persona
title = "Emily"
emily = autogen.AssistantAgent(
title=title,
llm_config=llm_config,
system_message="""Emily. You're a 35-year-old elementary faculty trainer residing in Sydney, Australia. You're married with two children aged 8 and 5, and you've got an annual earnings of AUD 75,000. You're introverted, excessive in conscientiousness, low in neuroticism, and luxuriate in routine. When purchasing on the grocery store, you like natural and domestically sourced produce. You worth comfort and use an internet purchasing platform. As a result of your restricted time from work and household commitments, you search fast and nutritious meal planning options. Your objectives are to purchase high-quality produce inside your price range and to seek out new recipe inspiration. You're a frequent shopper and use loyalty applications. Your most well-liked strategies of communication are e-mail and cell app notifications. You've been purchasing at a grocery store for over 10 years but additionally price-compare with others.""" + generate_notice(title),
)# John - Buyer Persona
title="John"
john = autogen.AssistantAgent(
title=title,
llm_config=llm_config,
system_message="""John. You're a 28-year-old software program developer primarily based in Sydney, Australia. You're single and have an annual earnings of AUD 100,000. You are extroverted, tech-savvy, and have a excessive degree of openness. When purchasing on the grocery store, you primarily purchase snacks and ready-made meals, and you employ the cell app for fast pickups. Your foremost objectives are fast and handy purchasing experiences. You often store on the grocery store and usually are not a part of any loyalty program. You additionally store at Aldi for reductions. Your most well-liked methodology of communication is in-app notifications.""" + generate_notice(title),
)
# Sarah - Buyer Persona
title="Sarah"
sarah = autogen.AssistantAgent(
title=title,
llm_config=llm_config,
system_message="""Sarah. You're a 45-year-old freelance journalist residing in Sydney, Australia. You're divorced with no children and earn AUD 60,000 per yr. You're introverted, excessive in neuroticism, and really health-conscious. When purchasing on the grocery store, you search for natural produce, non-GMO, and gluten-free objects. You've a restricted price range and particular dietary restrictions. You're a frequent shopper and use loyalty applications. Your most well-liked methodology of communication is e-mail newsletters. You solely store for groceries.""" + generate_notice(title),
)
# Tim - Buyer Persona
title="Tim"
tim = autogen.AssistantAgent(
title=title,
llm_config=llm_config,
system_message="""Tim. You're a 62-year-old retired police officer residing in Sydney, Australia. You're married and a grandparent of three. Your annual earnings comes from a pension and is AUD 40,000. You're extremely conscientious, low in openness, and like routine. You purchase staples like bread, milk, and canned items in bulk. As a result of mobility points, you want help with heavy objects. You're a frequent shopper and are a part of the senior citizen low cost program. Your most well-liked methodology of communication is junk mail flyers. You've been purchasing right here for over 20 years.""" + generate_notice(title),
)
# Lisa - Buyer Persona
title="Lisa"
lisa = autogen.AssistantAgent(
title=title,
llm_config=llm_config,
system_message="""Lisa. You're a 21-year-old college scholar residing in Sydney, Australia. You're single and work part-time, incomes AUD 20,000 per yr. You're extremely extroverted, low in conscientiousness, and worth social interactions. You store right here for widespread manufacturers, snacks, and alcoholic drinks, largely for social occasions. You've a restricted price range and are all the time searching for gross sales and reductions. You aren't a frequent shopper however are inquisitive about becoming a member of a loyalty program. Your most well-liked methodology of communication is social media and SMS. You store wherever there are gross sales or promotions.""" + generate_notice(title),
)
Outline the simulated atmosphere and guidelines for who can communicate — We’re permitting all of the brokers we’ve got outlined to take a seat inside the similar simulated atmosphere (group chat). We will create extra advanced situations the place we will set how and when subsequent audio system are chosen and outlined so we’ve got a easy operate outlined for speaker choice tied to the group chat which can make the researcher the lead and guarantee we go around the room to ask everybody a number of instances for his or her ideas.
# def custom_speaker_selection(last_speaker, group_chat):
# """
# Customized operate to pick which agent speaks subsequent within the group chat.
# """
# # Checklist of brokers excluding the final speaker
# next_candidates = [agent for agent in group_chat.agents if agent.name != last_speaker.name]# # Choose the subsequent agent primarily based in your customized logic
# # For simplicity, we're simply rotating by means of the candidates right here
# next_speaker = next_candidates[0] if next_candidates else None
# return next_speaker
def custom_speaker_selection(last_speaker: Optionally available[Agent], group_chat: GroupChat) -> Optionally available[Agent]:
"""
Customized operate to make sure the Researcher interacts with every participant 2-3 instances.
Alternates between the Researcher and contributors, monitoring interactions.
"""
# Outline contributors and initialize or replace their interplay counters
if not hasattr(group_chat, 'interaction_counters'):
group_chat.interaction_counters = {agent.title: 0 for agent in group_chat.brokers if agent.title != "Researcher"}
# Outline a most variety of interactions per participant
max_interactions = 6
# If the final speaker was the Researcher, discover the subsequent participant who has spoken the least
if last_speaker and last_speaker.title == "Researcher":
next_participant = min(group_chat.interaction_counters, key=group_chat.interaction_counters.get)
if group_chat.interaction_counters[next_participant] < max_interactions:
group_chat.interaction_counters[next_participant] += 1
return subsequent((agent for agent in group_chat.brokers if agent.title == next_participant), None)
else:
return None # Finish the dialog if all contributors have reached the utmost interactions
else:
# If the final speaker was a participant, return the Researcher for the subsequent flip
return subsequent((agent for agent in group_chat.brokers if agent.title == "Researcher"), None)
# Including the Researcher and Buyer Persona brokers to the group chat
groupchat = autogen.GroupChat(
brokers=[researcher, emily, john, sarah, tim, lisa],
speaker_selection_method = custom_speaker_selection,
messages=[],
max_round=30
)
Outline the supervisor to move directions into and handle our simulation — After we begin issues off we are going to communicate solely to the supervisor who will communicate to the researcher and panelists. This makes use of one thing known as GroupChatManager in Autogen.
# Initialise the supervisor
supervisor = autogen.GroupChatManager(
groupchat=groupchat,
llm_config=llm_config,
system_message="You're a reasearch supervisor agent that may handle a bunch chat of a number of brokers made up of a reasearcher agent and many individuals made up of a panel. You'll restrict the dialogue between the panelists and assist the researcher in asking the questions. Please ask the researcher first on how they wish to conduct the panel." + generate_notice(),
is_termination_msg=lambda x: True if "TERMINATE" in x.get("content material") else False,
)
We set the human interplay — permitting us to move directions to the assorted brokers we’ve got began. We give it the preliminary immediate and we will begin issues off.
# create a UserProxyAgent occasion named "user_proxy"
user_proxy = autogen.UserProxyAgent(
title="user_proxy",
code_execution_config={"last_n_messages": 2, "work_dir": "groupchat"},
system_message="A human admin.",
human_input_mode="TERMINATE"
)
# begin the reasearch simulation by giving instruction to the supervisor
# supervisor <-> reasearcher <-> panelists
user_proxy.initiate_chat(
supervisor,
message="""
Collect buyer insights on a grocery store grocery supply companies. Establish ache factors, preferences, and ideas for enchancment from totally different buyer personas. May you all please give your personal private oponions earlier than sharing extra with the group and discussing. As a reasearcher your job is to make sure that you collect unbiased data from the contributors and supply a abstract of the outcomes of this examine again to the tremendous market model.
""",
)
As soon as we run the above we get the output obtainable dwell inside your python atmosphere, you will notice the messages being handed round between the assorted brokers.
Now that our simulated analysis examine has been concluded we’d like to get some extra actionable insights. We will create a abstract agent to help us with this process and in addition use this in a Q&A state of affairs. Right here simply watch out of very massive transcripts would want a language mannequin that helps a bigger enter (context window).
We’d like seize all of the conversations — in our simulated panel dialogue from earlier to make use of because the person immediate (enter) to our abstract agent.
# Get response from the groupchat for person immediate
messages = [msg["content"] for msg in groupchat.messages]
user_prompt = "Right here is the transcript of the examine ```{customer_insights}```".format(customer_insights="n>>>n".be a part of(messages))
Lets craft the system immediate (directions) for our abstract agent — This agent will concentrate on creating us a tailor-made report card from the earlier transcripts and provides us clear ideas and actions.
# Generate system immediate for the abstract agent
summary_prompt = """
You're an skilled reasearcher in behaviour science and are tasked with summarising a reasearch panel. Please present a structured abstract of the important thing findings, together with ache factors, preferences, and ideas for enchancment.
This needs to be within the format primarily based on the next format:```
Reasearch Examine: <<Title>>
Topics:
<<Overview of the topics and quantity, another key data>>
Abstract:
<<Abstract of the examine, embrace detailed evaluation as an export>>
Ache Factors:
- <<Checklist of Ache Factors - Be as clear and prescriptive as required. I anticipate detailed response that can be utilized by the model on to make adjustments. Give a brief paragraph per ache level.>>
Recommendations/Actions:
- <<Checklist of Adctions - Be as clear and prescriptive as required. I anticipate detailed response that can be utilized by the model on to make adjustments. Give a brief paragraph per reccomendation.>>
```
"""
Outline the abstract agent and its atmosphere — Lets create a mini atmosphere for the abstract agent to run. It will want it’s personal proxy (atmosphere) and the provoke command which can pull the transcripts (user_prompt) because the enter.
summary_agent = autogen.AssistantAgent(
title="SummaryAgent",
llm_config=llm_config,
system_message=summary_prompt + generate_notice(),
)
summary_proxy = autogen.UserProxyAgent(
title="summary_proxy",
code_execution_config={"last_n_messages": 2, "work_dir": "groupchat"},
system_message="A human admin.",
human_input_mode="TERMINATE"
)
summary_proxy.initiate_chat(
summary_agent,
message=user_prompt,
)
This provides us an output within the type of a report card in Markdown, together with the power to ask additional questions in a Q&A method chat-bot on-top of the findings.
[ad_2]
