Structuring Inputs & Outputs in Multi Agent techniques Utilizing CrewAI

Instance of Pydantic Fashions

Let’s make a UserModel that inherits from BaseModel from Pydantic and the instantiated class ought to have an “id” as int, title as “str” and e mail as “e mail” 

from pydantic import BaseModel
class UserModel(BaseModel):
   id: int
   title: str
   e mail: str
# Legitimate enter
valid_user = UserModel(id=1, title="Vidhya", e mail="[email protected]")
print(valid_user)

id=1 title="Vidhya" e mail="[email protected]"

# Invalid enter (raises a validation error)
attempt:
   invalid_user = UserModel(id="one", title="Vidhya", e mail="[email protected]")
besides ValueError as e:
   print(f"Validation Error: {e}")

Validation Error: 1 validation error for UserModel
id
  Enter ought to be a legitimate integer, unable to parse string as an integer
[type=int_parsing, input_value="one", input_type=str]
    For additional info go to https://errors.pydantic.dev/2.9/v/int_parsing

We acquired a validation error once we tried passing a string in id and Pydantic additionally returns the hyperlink to documentation that helps us perceive the error.

Now, let’s take a look at different options of Pydantic, equivalent to non-compulsory, date, and default worth. 

from pydantic import BaseModel
from typing import Non-compulsory
from datetime import date
class EventModel(BaseModel):
   event_name: Non-compulsory[str] = None  # It is Non-compulsory to supply a price
   event_loc: str = "India" # Default worth
   event_date: date
# Computerized conversion of a string to a `date` object
occasion = EventModel(event_date="2024-10-17")
print(occasion)

event_name=None event_loc="India" event_date=datetime.date(2024, 10, 17)

We’ve handed solely the event_date, but it’s not throwing any error as a result of we set event_name to Non-compulsory and event_loc to “India” by default. Additionally, discover that the date is being typecast from string to datetime. 

Installations

We will probably be utilizing CrewAI all through this information, so be sure you set up CrewAI.

!pip set up crewai

Structuring Inputs in Agentic Programs

The inputs will be formatted whereas defining the Agent and Job utilizing curly braces { } with a variable title inside. Setting the human_input=True will immediate the person for suggestions for the output. Let’s outline an agent and job to reply questions associated to physics:

Be aware: I’ll be utilizing the  ‘gpt-4o-mini-2024-07-18’ from openai all through the information. 

from crewai import Agent, Job, Crew
import os
os.environ['OPENAI_API_KEY']=''
llm_config={'mannequin':'gpt-4o-mini-2024-07-18'}
teacher_agent = Agent(
 position="Physics Professor",
 purpose="You've gotten sturdy foundational data on physics "
       "You give the most effective reply to your college students",
 backstory=(
   "You're employed as a school professor"
   "You educate physics and make clear doubts {doubts}"
   "You reply the query solely whether it is associated to physics"
   "You just be sure you present the most effective solutions!"
 ),
 allow_delegation=False,
 verbose=True)
doubt_clarification_task = Job(
   description=(
       "Deal with the physics doubts raised by the scholar {doubts}. "
       "You do not reply the query if the query will not be associated to physics"
       "Present clear, detailed explanations, examples, and any mandatory formulation to assist the scholar perceive the ideas."
   ),
   expected_output=(
       "A transparent response to the scholar's doubts, "
       "together with explanations and examples if wanted. "
   ),
   agent=teacher_agent,
   human_input=True
)
crew = Crew(
 brokers=[teacher_agent],
 duties=[doubt_clarification_task],
 verbose=True,
 reminiscence=False
)

The inputs will be handed to the “inputs “ parameter in crew.kickoff()

inputs = {
   "doubts": "What's thermodynamics",
}
consequence = crew.kickoff(inputs=inputs)

Output

# Agent: Physics Professor
## Job: Deal with the physics doubts raised by the scholar What's
thermodynamics. You do not reply the query if the query will not be associated
to physicsProvide clear, detailed explanations, examples, and any mandatory
formulation to assist the scholar perceive the ideas.
# Agent: Physics Professor
## Ultimate Reply: 
Thermodynamics is the department of physics that offers with the relationships
between warmth, work, temperature, and vitality. It supplies a framework for
understanding how vitality is transferred and remodeled in techniques, whether or not
they're mechanical engines, fridges, or organic organisms.
The examine of thermodynamics is predicated on 4 elementary legal guidelines that describe
how vitality strikes and modifications type:
1. **Zeroth Legislation of Thermodynamics**: This regulation establishes the idea of 
temperature. It states that if two techniques are in thermal equilibrium with a
third system, then they're additionally in thermal equilibrium with one another. For
instance, if System A is in equilibrium with System C, and System B can also be
in equilibrium with System C, then System A and System B are in equilibrium 
with one another. This enables us to outline temperature as a measurable entity.
2. **First Legislation of Thermodynamics**: Typically said because the regulation of vitality 
conservation, this regulation asserts that vitality can't be created or destroyed,
solely remodeled. Mathematically, it may be expressed as:
   [
   Delta U = Q - W
   ]
   the place ( Delta U ) is the change in inner vitality of the system, ( Q
) is the warmth added to the system, and ( W ) is the work carried out by the 
system. An instance of the primary regulation in follow is a steam engine, the place warmth
vitality is transformed into work.
3. **Second Legislation of Thermodynamics**: This regulation introduces the idea of
entropy, stating that in any vitality switch, the whole entropy of an
remoted system can by no means lower over time. It means that vitality
transformations will not be 100% environment friendly and that techniques naturally progress
towards a state of dysfunction. A typical instance is the concept that warmth flows
spontaneously from a warmer physique to a colder physique, not the reverse.
4. **Third Legislation of Thermodynamics**: This regulation states that because the temperature
of a system approaches absolute zero, the entropy of an ideal crystal
approaches zero. This means that it's unimaginable to succeed in absolute zero
in a finite variety of steps, an idea that underscores the basic
limits of thermodynamic processes.
In sensible purposes, thermodynamics is utilized in quite a few fields equivalent to
engineering (designing engines, fridges), chemistry (response
spontaneity), and environmental science (understanding local weather techniques). 
In abstract, thermodynamics is crucial for analyzing how vitality is utilized
and transferred, offering insights into enhancing effectivity and
understanding pure processes in varied bodily techniques.
 ## Ultimate Outcome: Thermodynamics is the department of physics that offers with the
relationships between warmth, work, temperature, and vitality. It supplies a
framework for understanding how vitality is transferred and remodeled in
techniques, whether or not they're mechanical engines, fridges, or organic
organisms.
The examine of thermodynamics is predicated on 4 elementary legal guidelines that describe
how vitality strikes and modifications type:
1. **Zeroth Legislation of Thermodynamics**: This regulation establishes the idea of
temperature. It states that if two techniques are in thermal equilibrium with a
third system, then they're additionally in thermal equilibrium with one another. For
instance, if System A is in equilibrium with System C, and System B can also be
in equilibrium with System C, then System A and System B are in equilibrium
with one another. This enables us to outline temperature as a measurable
entity.
2. **First Legislation of Thermodynamics**: Typically said because the regulation of vitality
conservation, this regulation asserts that vitality can't be created or destroyed,
solely remodeled. Mathematically, it may be expressed as:
   [
   Delta U = Q - W
   ]
   the place ( Delta U ) is the change in inner vitality of the system, ( Q
) is the warmth added to the system, and ( W ) is the work carried out by the
system. An instance of the primary regulation in follow is a steam engine, the place
warmth vitality is transformed into work.
3. **Second Legislation of Thermodynamics**: This regulation introduces the idea of
entropy, stating that in any vitality switch, the whole entropy of an
remoted system can by no means lower over time. It means that vitality
transformations will not be 100% environment friendly and that techniques naturally progress
towards a state of dysfunction. A typical instance is the concept that warmth flows
spontaneously from a warmer physique to a colder physique, not the reverse.
4. **Third Legislation of Thermodynamics**: This regulation states that because the temperature
of a system approaches absolute zero, the entropy of an ideal crystal
approaches zero. This means that it's unimaginable to succeed in absolute zero
in a finite variety of steps, an idea that underscores the basic
limits of thermodynamic processes.
In sensible purposes, thermodynamics is utilized in quite a few fields equivalent to
engineering (designing engines, fridges), chemistry (response
spontaneity), and environmental science (understanding local weather techniques). 
In abstract, thermodynamics is crucial for analyzing how vitality is utilized
and transferred, offering insights into enhancing effectivity and 
understanding pure processes in varied bodily techniques.
=====
## Please present suggestions on the Ultimate Outcome and the Agent's actions:
Give the reply in 3 traces
# Agent: Physics Professor
## Ultimate Reply: 
Thermodynamics is the department of physics that offers with the relationships
between warmth, work, temperature, and vitality. It's ruled by 4
elementary legal guidelines that describe how vitality is transferred and transformed in
bodily techniques. For instance, the primary regulation of thermodynamics states that
vitality can't be created or destroyed, solely remodeled, which is expressed
mathematically as ΔU = Q - W, the place ΔU is the change in inner vitality, Q
is the warmth added to the system, and W is the work carried out by the system.

We handed a query within the enter, “What’s thermodynamics?” and with the assistance of a human enter, we acquired the specified reply. 

Structuring Outputs in Agentic Programs

Let’s make brokers who may help me fill out particulars like title, e mail, telephone quantity, and job one after the other. Structuring the outputs as Pydantic or json helps outline what output we’re anticipating and the format we count on in order that the next brokers get structured knowledge and context.

Be aware: In a case like mine, not structuring the outputs with anticipated fields would possibly result in a lack of info, and the next brokers would possibly begin hallucinating. 

from crewai import Agent, Job, Crew
import os
os.environ['OPENAI_API_KEY']=''
llm_config={'mannequin':'gpt-4o-mini-2024-07-18'}

Let’s outline the courses utilizing pedantic with the fields I count on within the output. As defined within the Pydantic fashions’ part, you need to use Non-compulsory or set a price by default if wanted.

from pydantic import BaseModel
from typing import Listing
# Outline the courses utilizing Pydantic
class nameEmail(BaseModel):
   title: Listing[str]
   e mail: Listing[str]
class phoneNo(BaseModel):
   title: str
   e mail: str
   telephone: int
class allDetails(BaseModel):
   title: str
   e mail: Listing[str]
   telephone: int
   job: str

Let’s outline the primary agent and job that fills out the person’s title and e mail. Use the “output_pydantic” parameter in Job(). I handed the nameEmail class as I gained’t be utilizing title and e mail as output right here. 

# Job 1: Agent for filling title and e mail utilizing output_pydantic
name_email_agent = Agent(
   position="Knowledge Entry Specialist",
   purpose="Your job is to fill out the person's title and e mail by yourself.",
   backstory=(
       "You focus on filling out private info like title and e mail."
       "You fill the title and e mail by yourself"
       "You employ uncommon names and use start yr within the e mail"
   ),
   allow_delegation=False,
   verbose=True
)
name_email_task = Job(
   description="Fill out the title and e mail of the person.",
   expected_output="Identify and e mail.",
   agent=name_email_agent,
   output_pydantic=nameEmail,
   human_input=False
)

Let’s use the opposite solution to construction the output through the use of the “output_json” parameter in Job() which provides the output in json format. 

# Job 2: Agent for filling telephone quantity utilizing output_json
phone_number_agent = Agent(
   position="Contact Info Supervisor",
   purpose="Your job is to fill out a random 10 digit telephone quantity by yourself.",
   backstory=(
       "You're an skilled in managing contact particulars, particularly telephone numbers."
   ),
   allow_delegation=False,
   verbose=True
)
phone_number_task = Job(
   description="Fill out the person's telephone quantity.",
   expected_output="A JSON format with the person's telephone quantity.",
   agent=phone_number_agent,
   output_json=phoneNo,
   human_input=False
)

Right here, the agent is designed to fill out the job particulars for the person. The ultimate output will comprise all of the title, e mail, telephone quantity, and job position info as pedantic and saved in output.txt utilizing the “output_file” parameter. 

# Job 3: Agent for filling job description utilizing output_file
job_description_agent = Agent(
   position="Job Decider",
   purpose="Your job is to randomly determine a job for the person",
   backstory=(
       "You deal with assigning job roles randomly."
   ),
   allow_delegation=False,
   verbose=True
)

job_description_task = Job(

   description="Fill out the person's job description.",

   expected_output="Show title, e mail, telephone no and job description.",

   agent=job_description_agent,

   output_pydantic=allDetails,

   output_file="/content material/output.txt",

   human_input=False

)
# Outline the crew to run all duties
crew = Crew(
   brokers=[name_email_agent, phone_number_agent, job_description_agent],
   duties=[name_email_task, phone_number_task, job_description_task],
   verbose=True,
   reminiscence=False
)
consequence = crew.kickoff()

# Agent: Knowledge Entry Specialist
## Job: Fill out the title and e mail of the person.
# Agent: Knowledge Entry Specialist
## Ultimate Reply: 
Identify: Elowen Thorne  
Electronic mail: [email protected]
# Agent: Contact Info Supervisor
## Job: Fill out the person's telephone quantity.
# Agent: Contact Info Supervisor
## Ultimate Reply: 
{
  "title": "Elowen Thorne",
  "e mail": "[email protected]",
  "phone_number": "1234567890"
}
# Agent: Job Decider
## Job: Fill out the person's job position.
# Agent: Job Decider
## Ultimate Reply: 
Identify: Elowen Thorne  
Electronic mail: [email protected]  
Cellphone No: 1234567890  
Job Position: Knowledge Analyst

That is the output file that was created :

You can even take a look at every job’s output:

job_description_task.output

TaskOutput(description="Fill out the person's job position.", title=None,
expected_output="Show title, e mail, telephone no and job position.", abstract="Fill
out the person's job position....", uncooked='Identify: Elowen Thorne  nEmail: 
[email protected]  nPhone No: 1234567890  nJob Position: Knowledge
Analyst', pydantic=allDetails(title="Elowen Thorne", e mail=
['[email protected]'], telephone=1234567890, job='Knowledge Analyst'),
json_dict=None, agent="Job Decider", output_format= 'pydantic'>)

I acquired the anticipated output, and as a result of structuring the intermediate outputs, the main points remained fixed as every job progressed. 

Additionally, to grasp the Agent AI higher, discover: The Agentic AI Pioneer Program.

This text explored the significance of structuring inputs and outputs in multi-agent utilizing instruments like Pydantic fashions and CrewAI. By organizing knowledge successfully and validating inputs, we will considerably improve the efficiency and reliability of multi-agent techniques. Structured outputs assist keep consistency and stop errors equivalent to knowledge loss and hallucinations, guaranteeing that brokers can collaborate effectively. Implementing these methods permits builders to create extra sturdy agentic techniques able to dealing with complicated duties with higher precision and reliability.

Ceaselessly Requested Questions

We use cookies important for this web site to operate properly. Please click on to assist us enhance its usefulness with extra cookies. Study our use of cookies in our Privateness Coverage & Cookies Coverage.

Present particulars

Settle for all cookies

Use mandatory cookies

Cookies

This web site makes use of cookies to make sure that you get the most effective expertise attainable. To be taught extra about how we use cookies, please check with our Privateness Coverage & Cookies Coverage.

brahmaid

It’s wanted for personalizing the web site.

csrftoken

This cookie is used to forestall Cross-site request forgery (typically abbreviated as CSRF) assaults of the web site

Identityid

Preserves the login/logout state of customers throughout the entire web site.

sessionid

Preserves customers’ states throughout web page requests.

g_state

Google One-Faucet login provides this g_state cookie to set the person standing on how they work together with the One-Faucet modal.

MUID

Utilized by Microsoft Readability, to retailer and monitor visits throughout web sites.

_clck

Utilized by Microsoft Readability, Persists the Readability Person ID and preferences, distinctive to that web site, on the browser. This ensures that habits in subsequent visits to the identical web site will probably be attributed to the identical person ID.

_clsk

Utilized by Microsoft Readability, Connects a number of web page views by a person right into a single Readability session recording.

SRM_I

Collects person knowledge is particularly tailored to the person or machine. The person will also be adopted outdoors of the loaded web site, creating an image of the customer’s habits.

SM

Use to measure the usage of the web site for inner analytics

CLID

The cookie is ready by embedded Microsoft Readability scripts. The aim of this cookie is for heatmap and session recording.

SRM_B

Collected person knowledge is particularly tailored to the person or machine. The person will also be adopted outdoors of the loaded web site, creating an image of the customer’s habits.

_gid

This cookie is put in by Google Analytics. The cookie is used to retailer info of how guests use a web site and helps in creating an analytics report of how the web site is doing. The info collected contains the variety of guests, the supply the place they’ve come from, and the pages visited in an nameless type.

_ga_#

Utilized by Google Analytics, to retailer and rely pageviews.

_gat_#

Utilized by Google Analytics to gather knowledge on the variety of occasions a person has visited the web site in addition to dates for the primary and most up-to-date go to.

acquire

Used to ship knowledge to Google Analytics in regards to the customer’s machine and habits. Tracks the customer throughout units and advertising and marketing channels.

AEC

cookies be certain that requests inside a looking session are made by the person, and never by different websites.

G_ENABLED_IDPS

use the cookie when clients need to make a referral from their gmail contacts; it helps auth the gmail account.

test_cookie

This cookie is ready by DoubleClick (which is owned by Google) to find out if the web site customer’s browser helps cookies.

_we_us

that is used to ship push notification utilizing webengage.

WebKlipperAuth

utilized by webenage to trace auth of webenagage.

ln_or

Linkedin units this cookie to registers statistical knowledge on customers’ habits on the web site for inner analytics.

JSESSIONID

Use to keep up an nameless person session by the server.

li_rm

Used as a part of the LinkedIn Bear in mind Me function and is ready when a person clicks Bear in mind Me on the machine to make it simpler for her or him to register to that machine.

AnalyticsSyncHistory

Used to retailer details about the time a sync with the lms_analytics cookie came about for customers within the Designated International locations.

lms_analytics

Used to retailer details about the time a sync with the AnalyticsSyncHistory cookie came about for customers within the Designated International locations.

liap

Cookie used for Signal-in with Linkedin and/or to permit for the Linkedin comply with function.

go to

permit for the Linkedin comply with function.

li_at

typically used to determine you, together with your title, pursuits, and former exercise.

s_plt

Tracks the time that the earlier web page took to load

lang

Used to recollect a person’s language setting to make sure LinkedIn.com shows within the language chosen by the person of their settings

s_tp

Tracks p.c of web page considered

AMCV_14215E3D5995C57C0A495C55percent40AdobeOrg

Signifies the beginning of a session for Adobe Expertise Cloud

s_pltp

Supplies web page title worth (URL) to be used by Adobe Analytics

s_tslv

Used to retain and fetch time since final go to in Adobe Analytics

li_theme

Remembers a person’s show choice/theme setting

li_theme_set

Remembers which customers have up to date their show / theme preferences

We don’t use cookies of this sort.

_gcl_au

Utilized by Google Adsense, to retailer and monitor conversions.

SID

Save sure preferences, for instance the variety of search outcomes per web page or activation of the SafeSearch Filter. Adjusts the advertisements that seem in Google Search.

SAPISID

Save sure preferences, for instance the variety of search outcomes per web page or activation of the SafeSearch Filter. Adjusts the advertisements that seem in Google Search.

__Secure-#

Save sure preferences, for instance the variety of search outcomes per web page or activation of the SafeSearch Filter. Adjusts the advertisements that seem in Google Search.

APISID

Save sure preferences, for instance the variety of search outcomes per web page or activation of the SafeSearch Filter. Adjusts the advertisements that seem in Google Search.

SSID

Save sure preferences, for instance the variety of search outcomes per web page or activation of the SafeSearch Filter. Adjusts the advertisements that seem in Google Search.

HSID

Save sure preferences, for instance the variety of search outcomes per web page or activation of the SafeSearch Filter. Adjusts the advertisements that seem in Google Search.

DV

These cookies are used for the aim of focused promoting.

NID

These cookies are used for the aim of focused promoting.

1P_JAR

These cookies are used to collect web site statistics, and monitor conversion charges.

OTZ

Mixture evaluation of web site guests

_fbp

This cookie is ready by Fb to ship ads when they’re on Fb or a digital platform powered by Fb promoting after visiting this web site.

fr

Comprises a novel browser and person ID, used for focused promoting.

bscookie

Utilized by LinkedIn to trace the usage of embedded companies.

lidc

Utilized by LinkedIn for monitoring the usage of embedded companies.

bcookie

Utilized by LinkedIn to trace the usage of embedded companies.

aam_uuid

Use these cookies to assign a novel ID when customers go to a web site.

UserMatchHistory

These cookies are set by LinkedIn for promoting functions, together with: monitoring guests in order that extra related advertisements will be offered, permitting customers to make use of the ‘Apply with LinkedIn’ or the ‘Signal-in with LinkedIn’ features, accumulating details about how guests use the location, and so forth.

li_sugr

Used to make a probabilistic match of a person’s identification outdoors the Designated International locations

MR

Used to gather info for analytics functions.

ANONCHK

Used to retailer session ID for a customers session to make sure that clicks from adverts on the Bing search engine are verified for reporting functions and for personalisation

We don’t use cookies of this sort.

Cookie declaration final up to date on 24/03/2023 by Analytics Vidhya.

Cookies are small textual content information that can be utilized by web sites to make a person’s expertise extra environment friendly. The regulation states that we will retailer cookies in your machine if they’re strictly mandatory for the operation of this web site. For all different sorts of cookies, we want your permission. This web site makes use of several types of cookies. Some cookies are positioned by third-party companies that seem on our pages. Be taught extra about who we’re, how one can contact us, and the way we course of private knowledge in our Privateness Coverage.

GenAI
Pinnacle
Program

Revolutionizing AI Studying & Improvement

• 1:1 Mentorship with Generative AI consultants
• Superior Curriculum with 200+ Hours of Studying
• Grasp 26+ GenAI Instruments and Libraries

Enter e mail tackle to proceed

Resend OTP

Resend OTP in 45s