tinytraitlet

tinytraitlet is a simple library that enables attribute validation and type checking for Traitful classes, inspired by traitlets library.

Usage

from tinytraitlet import Traitful, String, validate, TraitError


class User(Traitful):
    name = String()
   
    @validate('name')
    def validate_name(self, value):
        if value == 'undefined':
            raise TraitError("Name cannot be 'undefined'")


class SubscriptionUser(User):
    subscription = String()
   
    @validate('subscription')
    def validate_subscription(self, value):
        if value not in ["monthly", "yearly"]:
            raise TraitError("subscription must be either 'monthly' or 'yearly'")

These classes behave as one would expect. @validate('name') decorator registers the validate_name method as the validator for the name trait. SubscriptionUser inherits both the name trait and the validate_name validator.

>>> user = User(name="Niki")
>>> user.name
'Niki'
>>> user = SubscriptionUser(name="undefined")
tinytraitlet.TraitError: Name cannot be 'undefined'
>>> user = SubscriptionUser(subscription="yearly")
>>> user.subscription
'yearly'
>>> user.subscription = "daily"
tinytraitlet.TraitError: subscription must be either 'monthly' or 'yearly'

Defining custom types: We should define default_value and model_type attributes. Example:

from tinytraitlet import Model

class Integer(Model):
    default_value = 0
    model_type = int

Why solve the same problem?

tinytraitlet does not use metaclasses unlike traitlets. This reduces the chance of metaclass conflict. tinytraitlet is a simple alternative, currently under 100 lines of code, and offers only the absolutely essential functionality.

Limitations

• Currently a single trait type is defined: String.
• Default values are not validated.
• Cross validation is not supported.
• A trait cannot have multiple validators