Trust the Shape: Schema Validation & Negative Testing

The idea, in one line

Instead of poking at fields one by one, describe the shape the whole response should have, then check the reply against it in a single line. Pydantic v2 does this for you.

How to spot when you need it

• A key got renamed and your test never noticed
• A number arrived as a string ("123" instead of 123)
• A field came back null where you expected text
• You're writing five asserts to check one response body

See it work

You define a BaseModel with typed fields. Required fields have no default; optional ones use Optional[str] = None. Feed the response to Model.model_validate(data) (or model_validate_json for raw JSON). If every field fits, you get a clean typed object back. If anything is off, Pydantic raises a ValidationError that lists all the problems at once.

import requests
from pydantic import BaseModel
from typing import Optional

class Post(BaseModel):
    id: int
    userId: int
    title: str
    body: Optional[str] = None   # optional field

BASE = "https://jsonplaceholder.typicode.com"

def test_response_matches_schema():
    r = requests.get(f"{BASE}/posts/1", timeout=5)
    post = Post.model_validate(r.json())  # raises if shape is wrong
    assert post.id == 1

Read it top to bottom: you declared what a Post should look like, then handed the response to the model. One line replaced a fistful of field checks.

Advanced — negative testing and strict mode

Pydantic is helpful and will coerce sensibly — the string "123" becomes the int 123. When a type must be exact, turn on strict mode so it won't quietly convert. Schema checks also pair naturally with negative testing: the paths where things should fail on purpose.

• Send a bad payload, expect 400
• Drop the auth header, expect 401
• Ask for a missing id, expect 404

Grounded in the official Pydantic v2 docs (Models & validation)