Drafts

global Field
global BaseModel
from pydantic import BaseModel
from pydantic import Field

Pydantic is a Python library for serializing data into models that can be validated with a deep set of built in valitators or your own custom validators, and deserialize back to JSON or dictionary.

Installation #

To install pydantic you will first need python and pip. Once you have pip installed you can install pydantic with pip.

pip install pydantic

Always install in a virtual environment

Creating a Pydantic model #

To get started with pydantic you will first need to create a Pydantic model. This is a python class that inherits from pydantic.BaseModel.

from pydantic import BaseModel
from pydantic import Field
from typing import Optional

class Person(BaseModel):
    name: str = Field(...)
    age: int

parsing an object #

person = Person(name="John Doe", age=30)
print(person)

name='John Doe' age=30

data serialization #

Pydantic has some very robust serialization methods that will automatically coherse your data into the type specified by the type-hint in the model if it can.

person = Person(name=12, age="30")
print(f'name: {person.name}, type: {type(person.name)}')
print(f'age: {person.age}, type: {type(person.age)}')

1 validation error for Person
name
  Input should be a valid string [type=string_type, input_value=12, input_type=int]
    For further information visit https://errors.pydantic.dev/2.3/v/string_type

person = Person(name="John Doe", age='thirty')
print(f'name: {person.name}, type: {type(person.name)}')
print(f'age: {person.age}, type: {type(person.age)}')

1 validation error for Person
age
  Input should be a valid integer, unable to parse string as an integer [type=int_parsing, input_value='thirty', input_type=str]
    For further information visit https://errors.pydantic.dev/2.3/v/int_parsing

loading from json #

serializing to json #

validation #

from kedro.pipeline import node

node(
    input="raw",
    output="int",
    func=my_func,
    tags=["one"],
)