README.rst

python-modeled

>>> import modeled
>>> print(modeled.__version__)
0.1

>>> print(modeled.__description__)
Universal data modeling for Python.

• Create classes with typed data members in pythonic style.
• Use typed containers.
• Map modeled classes and their instances to any kind of data storage, serialization, visualization or user interface by using existing adapters or writing custom ones.

https://bitbucket.org/userzimmermann/python-modeled

https://github.com/userzimmermann/python-modeled

0. Setup

Supported Python versions: 2.7, 3.3, 3.4

Just install the latest release with pip. It automatically installs all requirements:

pip install modeled

>>> modeled.__requires__
six
path.py>=7.0
moretools>=0.1.5

1. Writing modeled classes

All Python classes are derived from object. Modeled classes are derived from modeled.object, which has a predefined mobject shortcut:

from modeled import mobject

class MClass(mobject):
    # optional
    class model:
        name = 'NotMClass'

        group__option = 'value'

        class group:
            other_option = 'other value'

Adding modeled members

Modeled members are the typed attributes of modeled objects. They are defined as modeled class attributes and connect to class instances via Python's descriptor mechanism. These member descriptors are instances of the modeled.member base class, which has a predefined m shortcut. Typed member subclasses are created by passing any type (class object) in [...] brackets to the member base class:

from modeled import mobject, m

class MClass(mobject):
    some_int = m[int]
    some_float = m[float]
    some_string = m[str]

These typed member suclasses are created only once for each given type. Modeled member descriptors are automatically instantiated by modeled.object's metaclass. You can easily access these instances on modeled class level:

>>> MClass.some_int
modeled.member[int]()

>>> MClass.some_int.mtype
int

Manually instantiate a member descriptor for defining a default value or passing extra keyword arguments or options:

class MClass(mobject):
    some_int = m[int](4, name='not_some_int', group__option='value')

Other keyword arguments are:

• new= Any callable object which will be called instead of member's data type if setting a member value which is not an instance of data type. Must return a data type instance.
• choices= A sequence of allowed member values.
• changed= A sequence of callback functions (any callable objects) to be triggered after setting member values. Get the modeled class instance as first and the value as second argument.

When providing a default value, the data type can be omitted:

>>> m(4)
modeled.member[int](4)