This documentation is for a development version of Traitlets. There may be significant differences from the latest stable release.

Using Traitlets

In short, traitlets let the user define classes that have

  1. Attributes (traits) with type checking and dynamically computed default values
  2. Traits emit change events when attributes are modified
  3. Traitlets perform some validation and allow coercion of new trait values on assignment. They also allow the user to define custom validation logic for attributes based on the value of other attributes.

Default values, and checking type and value

At its most basic, traitlets provides type checking, and dynamic default value generation of attributes on :class:traitlets.HasTraits subclasses:

import getpass

class Identity(HasTraits):
    username = Unicode()

    def _default_username(self):
        return getpass.getuser()
class Foo(HasTraits):
    bar = Int()

foo = Foo(bar='3')  # raises a TraitError
TraitError: The 'bar' trait of a Foo instance must be an int,
but a value of '3' <class 'str'> was specified


Traitlets implement the observer pattern

class Foo(HasTraits):
    bar = Int()
    baz = Unicode()

foo = Foo()

def func(change):
    print(change['new'])   # as of traitlets 4.3, one should be able to
                           # write print( instead

foo.observe(func, names=['bar']) = 1  # prints '0\n 1'
foo.baz = 'abc'  # prints nothing

When observers are methods of the class, a decorator syntax can be used.

class Foo(HasTraits):
    bar = Int()
    baz = Unicode()

    def _observe_bar(self, change):

Validation and Coercion

Custom Cross-Validation

Each trait type (Int, Unicode, Dict etc.) may have its own validation or coercion logic. In addition, we can register custom cross-validators that may depend on the state of other attributes.

Basic Example: Validating the Parity of a Trait

from traitlets import HasTraits, TraitError, Int, Bool, validate

class Parity(HasTraits):
    value = Int()
    parity = Int()

    def _valid_value(self, proposal):
        if proposal['value'] % 2 != self.parity:
            raise TraitError('value and parity should be consistent')
        return proposal['value']

    def _valid_parity(self, proposal):
        parity = proposal['value']
        if parity not in [0, 1]:
            raise TraitError('parity should be 0 or 1')
        if self.value % 2 != parity:
            raise TraitError('value and parity should be consistent')
        return proposal['value']

parity_check = Parity(value=2)

# Changing required parity and value together while holding cross validation
with parity_check.hold_trait_notifications():
    parity_check.value = 1
    parity_check.parity = 1

However, we recommend that custom cross-validators don’t modify the state of the HasTraits instance.

Advanced Example: Validating the Schema

The List and Dict trait types allow the validation of nested properties.

from traitlets import HasTraits, Dict, Bool, Unicode

class Nested(HasTraits):

    value = Dict(traits={
        'configuration': Dict(trait=Unicode()),
        'flag': Bool()

n = Nested()
n.value = dict(flag=True, configuration={})  # OK
n.value = dict(flag=True, configuration='')  # raises a TraitError.

However, for deeply nested properties it might be more appropriate to use an external validator:

import jsonschema

value_schema = {
     'type' : 'object',
     'properties' : {
         'price' : { 'type' : 'number' },
         'name' : { 'type' : 'string' },

from traitlets import HasTraits, Dict, TraitError, validate, default

class Schema(HasTraits):

    value = Dict()

    def _default_value(self):
        return dict(name='', price=1)

    def _validate_value(self, proposal):
            jsonschema.validate(proposal['value'], value_schema)
        except jsonschema.ValidationError as e:
            raise TraitError(e)
        return proposal['value']

s = Schema()
s.value = dict(name='', price='1')  # raises a TraitError

Holding Trait Cross-Validation and Notifications

Sometimes if may be impossible to transition from to valid states for a HasTraits instance by change attributes one by one. The hold_trait_notifications context manager can be used to hold the custom cross validation until the context manager is released. If a validation error occurs, changes are rolled back to the initial state.

Custom Events

Finally, trait types can emit other events types than trait changes. This capability was added so as to enable notifications on change of values in container classes. The items available in the dictionary passed to the observer registered with observe depends on the event type.