pydantic validates your inputs, or does it?

TL;DR

🎊 pydantic calls itself the most widely used data validation library for Python

😧 Yet, it does not validate default values, nor assignments by default. That's ... unexpected

Background

pydantic is a Python library that tries to bring type-safety to Python. Rather than type hints (it may be true, but doesn't have to be), pydantic will validate an input. An example:

from pydantic import BaseModel
 
class Test(BaseModel):
    foo: str
 
t = Test(foo=[1,2,3])  # Raises a ValidationError

Let's have some fun with pydantic validation

However, this works without problem:  

from pydantic import BaseModel
 
class Test(BaseModel):
    foo: str = [1,2,3]

  Equally, this also works:  

from pydantic import BaseModel
 
class Test(BaseModel):
    foo: str
 
t = Test(foo="1")
t.foo = [1,2,3]

  Why? Because pydantic argues that this is performance plus backwards compatibility, at least according to this Github issue. I disagree. The reason to use pydantic is type-safety. And being able to violate the type-safety in the default settings  is not expected behavior.   Alas, there is a setting so solve it:  

class Test(BaseModel, validate_default=True, validate_assignment=True):
    foo: str

  As you would expect, the validate_default validates the default values, while validate_assignment validates the assignment.