valdec

Decorator for validating arguments and/or result of functions and methods of a class against annotations.

Can be used in synchronous or asynchronous version.

Validation is supported using pydantic.BaseModel and validated_dc.ValidatedDC.

You can easily add support for any other validator. To do this, you need to write your own validator-function and specify it in the settings for decorator (how to do this - see the examples below).

Note: if the result of the function has no annotation, the result is expected to be None.

Installation

pip install valdec

Quick example

Default

Based on the validator pydantic.BaseModel:

pip install pydantic

from typing import List, Optional

from pydantic import BaseModel, StrictInt, StrictStr
from valdec.decorators import validate
# from valdec.decorators import async_validate  # Use for async functions


@validate  # all arguments with annotations and return
def func(i: StrictInt, s: StrictStr) -> StrictInt:
    return i

assert func(1, "s") == 1


@validate("i", exclude=True)  # exclude "i" (only "s" and return)
def func(i: StrictInt, s: StrictStr) -> StrictInt:
    return int(i)

assert func("1", "s") == 1


# For example, an Api got a json-string and decoded it:
data_for_group = [
    {"name": "Peter", "profile": {"age": 22, "city": "Samara"}},
    {"name": "Elena", "profile": {"age": 20, "city": "Kazan"}},
]

class Profile(BaseModel):
    age: StrictInt
    city: StrictStr

class Student(BaseModel):
    name: StrictStr
    profile: Profile

@validate("s", "group")  # only "s" and "group"
def func(i: StrictInt, s: StrictStr, group: Optional[List[Student]] = None):

    assert i == "i not int"  # because "i" is not validated
    assert isinstance(s, str)
    for student in group:
        # `student` is now an instance of `Student`
        # (This conversion can be disabled!)
        assert isinstance(student, Student)
        assert isinstance(student.name, str)
        assert isinstance(student.profile.age, int)

    return group

result = func("i not int", "string",  data_for_group)
# The result can be any, because it is not validated.., now this is a list:
assert isinstance(result, list)

validated_dc_validator

It is possible to enable validation with validated_dc.ValidatedDC. To do this, change the reference to the validator-function in the decorator settings:

pip install validated-dc

from dataclasses import dataclass
from typing import List, Optional

from valdec.data_classes import Settings
from valdec.decorators import validate as _validate
from valdec.validator_validated_dc import validator 
from validated_dc import ValidatedDC

custom_settings = Settings(
    validator=validator,     # function for validation
    is_replace_args=True,    # default
    is_replace_result=True,  # default
    extra={}                 # default
)
def validate(*args, **kwargs):  # define new decorator
    kwargs["settings"] = custom_settings
    return _validate(*args, **kwargs)


# The new decorator can now be used:

@validate  # all arguments with annotations and return
def func(i: int, s: str) -> int:
    return i

assert func(1, "s") == 1


@validate("s")  # only "s"
def func(i: int, s: str) -> int:
    return i

assert func("not int", "s") == "not int"


@validate("s", "return")  # only "s" and return
def func(i: int, s: str) -> int:
    return int(i)

assert func("1", "s") == 1


@validate("i", exclude=True)  # exclude "i" (only "s" and return)
def func(i: int, s: str) -> int:
    return int(i)

assert func("1", "s") == 1


data_for_group = [
    {"name": "Peter", "profile": {"age": 22, "city": "Samara"}},
    {"name": "Elena", "profile": {"age": 20, "city": "Kazan"}},
]

@dataclass
class Profile(ValidatedDC):
    age: int
    city: str

@dataclass
class Student(ValidatedDC):
    name: str
    profile: Profile

@validate("s", "group")  # only "s" and "group"
def func(i: int, s: str, group: Optional[List[Student]] = None):

    assert i == "i not int"  # because "i" is not validated
    assert isinstance(s, str)
    for student in group:
        # `student` is now an instance of `Student`
        # (This behavior can be changed by is_replace_args=False in the
        # Settings instance)
        assert isinstance(student, Student)
        assert isinstance(student.name, str)
        assert isinstance(student.profile.age, int)

    return group

result = func("i not int", "string",  data_for_group)
# The result can be any, because it is not validated.., now this is a list:
assert isinstance(result, list)

Errors

ValidationArgumentsError is raised on an arguments validation error, and a ValidationReturnError on a result validation error:

from typing import Optional

from pydantic import StrictInt, StrictStr
from valdec.decorators import validate
from valdec.errors import ValidationArgumentsError, ValidationReturnError


class Foo:

    @validate  # all arguments with annotations and return
    def bar_1(self, i: StrictInt, s: Optional[StrictStr] = None):
        pass

    @validate("return")  # only "return"
    def bar_2(self, i: StrictInt, s: Optional[StrictStr] = None):
        return i


foo = Foo()

foo.bar_1(1, "2")  # ok

try:
    foo.bar_1(1, 2)
except ValidationArgumentsError as error:
    print(type(error), error, "\n")


foo.bar_2(None, 1)  # ok

try:
    foo.bar_2(1, 2)
except ValidationReturnError as error:
    print(type(error), error)