Skip to content
Sarj Sarj Sarj Standards

no-mocked-pydantic-value-object

Build real Pydantic models when test doubles supply model field data.

SARJ479 Rule source Tests Catalog JSON

Why

Mocking a Pydantic value object bypasses validation, defaults, serialization, and field relationships. A specced mock still permits test data that the real model could never produce. Real model fixtures keep those contracts active while allowing services and external boundaries to remain mocked.

Fix

Construct the real model, using a small typed factory for repeated valid defaults. Keep mock behavior at collaborator boundaries; a model method deliberately configured to raise remains a valid failure-injection test.

Examples

Before — flagged A mock bypasses the model's validated data contract
tests/test_profile.py
from unittest.mock import Mock
from pydantic import BaseModel




class Profile(BaseModel):
    language: str




def test_profile():
    profile = Mock(spec=Profile)
    profile.language = "ar"
After — preferred Real model fixtures retain validation and defaults
tests/test_profile.py
from pydantic import BaseModel




class Profile(BaseModel):
    language: str




def make_profile(*, language: str = "en") -> Profile:
    return Profile(language=language)




def test_profile():
    profile = make_profile(language="ar")
    assert profile.language == "ar"