Requests

Pydantic models are a great way to validate and serialize data for requests and responses. Pydantic is instrumental in many web frameworks and libraries, such as FastAPI, Django, Flask, and HTTPX.

httpx requests

httpx is an HTTP client for Python 3 with synchronous and asynchronous APIs. In the below example, we query the JSONPlaceholder API to get a user's data and validate it with a Pydantic model.

import httpx

from pydantic import BaseModel, EmailStr


class User(BaseModel):
    id: int
    name: str
    email: EmailStr


url = 'https://jsonplaceholder.typicode.com/users/1'

response = httpx.get(url)
response.raise_for_status()

user = User.model_validate(response.json())
print(repr(user))
#> User(id=1, name='Leanne Graham', email='[email protected]')

The [TypeAdapter][pydantic.type_adapter.TypeAdapter] tool from Pydantic often comes in quite handy when working with HTTP requests. Consider a similar example where we are validating a list of users:

from pprint import pprint

import httpx

from pydantic import BaseModel, EmailStr, TypeAdapter


class User(BaseModel):
    id: int
    name: str
    email: EmailStr


url = 'https://jsonplaceholder.typicode.com/users/'  # (1)!

response = httpx.get(url)
response.raise_for_status()

users_list_adapter = TypeAdapter(list[User])

users = users_list_adapter.validate_python(response.json())
pprint([u.name for u in users])
"""
['Leanne Graham',
 'Ervin Howell',
 'Clementine Bauch',
 'Patricia Lebsack',
 'Chelsey Dietrich',
 'Mrs. Dennis Schulist',
 'Kurtis Weissnat',
 'Nicholas Runolfsdottir V',
 'Glenna Reichert',
 'Clementina DuBuque']
"""

1. Note, we're querying the /users/ endpoint here to get a list of users.

!!! tip "Logfire integration" When you validate responses like this, a [ValidationError][pydantic_core.ValidationError] is often the first sign that an API you depend on has changed its response format. The useful questions at that point are what did the response actually contain, and when did this start: recording failed validations with Logfire answers both, since each failure is stored with the data that triggered it.