Python Data Classes
Python Data Classes support Python 2 and 3 with many human-friendly features.
This project has been tested in Python 2.7 and Python 3.4+.
It is not related to the dataclasses module introduced in Python 3.7.
Feature Support
Python Data Classes generally support
- untyped data, typed data, and typed data list
- arbitrary attribute names in the dictionary input
- no python attribute naming conflict or pollution
- lazy and auto object creation on-the-fly
- lazy loading of the dict input (default: enabled)
- synchronous updates of the dict input (default: disabled)
Getting Started
- Install the Python Data Classes
pip install pydataclasses
- Create the Python Data Classes
from pydataclasses import DataClass
class TT(DataClass):
attr = None
class ONCE(DataClass):
data = None
class TWICE(ONCE):
tear = TT
tears = [TT]
def __init__(self, *args, **kwargs):
super(TWICE, self).__init__(*args, **kwargs)
self.pair = TWICE
self.pairs = [TWICE]- Play with the Python Data Classes
old_data = 'hello'
new_data = 'world'
old_dict = {'data': old_data}
new_dict = {'data': new_data}
# lazy and auto object creation on-the-fly
twice = TWICE()
assert twice.pairs[1].tear.attr is None
twice.pairs[1].tear.attr = old_data
assert old_data == twice.pairs[1].tear.attr
# lazy loading of the dict input (default: enabled)
twice = TWICE(old_dict)
assert old_data == twice.data
assert old_dict == twice.__as_dict__(dict, 0)
# synchronous updates of the dict input (default: disabled)
assert old_dict != new_dict
twice = TWICE(old_dict, __sync__=True)
twice.data = new_data
assert old_dict == new_dictBest Practices
Please
- read and refer to
pydataclasses.utils.JSONDataas an example - put a tailored subclass between
pydataclassesand your project - write (or copy from
pydataclasses) as many tests as possible for this subclass
from pydataclasses import DataClass
class ProjectDataClass(DataClass):
def as_dict(self, dict_class=OrderedDict):
return self.__as_dict__(dict_class, 0)
