Python Essentials – Validation with Pydantic

Bank Teller Verification

Imagine you go to a bank to open an account:

1. You fill out a form with your details (name, age, balance, ID).

2. The bank teller checks the form:

   • Is your age a number?

   • Is your ID valid?

   • Is the initial deposit non-negative?

3. If something is wrong → the teller rejects it and asks you to fix it.

4. If everything is correct → the account is created successfully.

This is exactly what Pydantic does for Python data: it validates incoming data and ensures it is the right type, format, and structure.

---

🔹 What is Pydantic?

Pydantic is a Python library used for data validation and settings management.

• Ensures your data is correct type and format

• Automatically parses JSON into Python objects

• Raises clear errors if the data is invalid

• Perfect for FastAPI APIs and Agentic AI

---

🔹 Step 1: Install Pydantic

pip install pydantic

---

🔹 Step 2: Define a Data Model

We define schemas for the data we expect:

from pydantic import BaseModel, Field, ValidationError

class User(BaseModel):
    name: str
    age: int
    balance: float = 0.0

• BaseModel → Base class for all Pydantic models

• Field → optional, can set defaults or constraints

• Each attribute has a type

---

🔹 Step 3: Create & Validate Data

# Valid input
user1 = User(name="Anita", age=25, balance=100.50)
print(user1)

Output:

name='Anita' age=25 balance=100.5

Invalid input:

try:
    user2 = User(name="Ravi", age="twenty", balance=-50)
except ValidationError as e:
    print(e)

Output:

1 validation error for User
age
  value is not a valid integer (type=type_error.integer)

✅ Pydantic automatically catches mistakes

---

🔹 Step 4: Add Extra Validation

You can enforce rules like “age must be > 0”:

from pydantic import BaseModel, Field, validator

class User(BaseModel):
    name: str
    age: int = Field(..., gt=0)
    balance: float = Field(default=0.0, ge=0.0)  # balance >= 0

    @validator('name')
    def name_must_not_be_empty(cls, v):
        if not v.strip():
            raise ValueError('Name cannot be empty')
        return v

• gt=0 → greater than 0

• ge=0 → greater than or equal to 0

• @validator → custom logic

---

🔹 Step 5: Real-Time Agentic AI Example

Imagine an Agent API that receives a task request:

from pydantic import BaseModel, Field, ValidationError
from fastapi import FastAPI

app = FastAPI()

class TaskRequest(BaseModel):
    task_name: str
    priority: int = Field(..., ge=1, le=5)  # 1 to 5
    data_source: str

@app.post("/run_task")
def run_task(task: TaskRequest):
    return {"message": f"Running task: {task.task_name} with priority {task.priority}"}

Example Requests:

Valid POST:

{
  "task_name": "Fetch AI Articles",
  "priority": 3,
  "data_source": "Web"
}

✅ Works perfectly

Invalid POST:

{
  "task_name": "Fetch AI Articles",
  "priority": 10,
  "data_source": "Web"
}

❌ Returns:

{
  "detail": [
    {
      "loc": ["body","priority"],
      "msg": "ensure this value is less than or equal to 5",
      "type": "value_error.number.not_le"
    }
  ]
}

• Pydantic automatically validates the request

• FastAPI returns a structured error message

• Ensures agents never get invalid data

---

🔹 Step 6: Why Validation Matters in Agentic AI

• Ensures agent receives correct inputs

• Prevents runtime crashes

• Makes APIs secure, predictable, and production-ready

• Helps debug quickly if input is wrong

---

✅ Key Takeaways

ConceptAnalogyPython/Pydantic ExampleValidationBank teller checks formsBaseModel + type hints + validatorDefaults / ConstraintsMinimum deposit / ageField(..., ge=0)Automatic Error ResponseTeller rejects invalid formFastAPI + Pydantic auto error messagesData ParsingForm → structured recordJSON → Python object

---

Mental Model for Agentic AI:

• Agent receives input → must be structured and valid

• Pydantic ensures validity → like a bank teller checking forms

• Invalid inputs rejected automatically → agent doesn’t crash