YAML Model

Semi-transparent YAML serialization/deserialization. What does this mean? Well, much like every other serialization method ever, you create your "fields" that are to be serialized, give the serializer data, then hit the go button. The difference here is that this mostly happens under the covers of a regular Python object.

Examples

from datetime import datetime
from yaml_model import LoadOnAccess, Model

class User(Model):
    """ A basic user model """
    email = LoadOnAccess(default='')
    created_ts = LoadOnAccess(generate=lambda _: datetime.now())

    def __init__(self, slug):
        self.slug = slug
        super(User, self).__init__()

# Save a new user
new_user = User('my_username')
new_user.email = 'hey@there.com'
new_user.save()

# This will save a file './data/users/my_username.yaml' with content:
# email: hey@there.com
# created_ts: 2015-02-01 15:41:23.99072

# Load up that same user
load_user = User('my_username')

print(load_user.email)
# Will print hey@there.com

print(type(load_user.create_ts))
# Will print datetime.datetime

print(load_user.created_ts)
# Will print datetime.datetime(2015, 2, 1, 15, 41, 23, 99072)

Generators and defaults

For the LoadOnAccess field type, there are 2 kwargs that you can specify for "default" values. The difference between them is relatively minor. Namely, the generate kwarg takes precidence over the default kwarg, and the generate kwarg output is persisted to the model file, whereas the default value is not. Other than that, they behave the same.

Both values may be either a static value, or a callable. Callables will be called, and passed the model it's being called on as the first argument.

ModelReference differences

When adding a ModelReference field, these fields are considered to return a cat slug, or an instance of a Model.