metaframe

MetaClass infrastructure to intercept instance creation/initialization enabling modification of args/kwargs and instance.

Documentation

Read the full documentation at readthedocs.org:

• metaframe documentation

Python 2/3 Support

• Python 2.7
• Python 3.2/3.3/3.4/3.5
• It also works with pypy and pypy3

Installation

metaframe is self-contained with no external dependencies

From pypi:

pip install metaframe

From source:

• Place the metaframe directory found in the sources inside your project

Features:

• MetaFrame metaclass to apply to any object - With embedded staticmethod with_metaclass to enable inheritance
• MetaFrameBase class from which classes can inherit
• 3 hooks (classmethods)
  • _new_pre: called before object creation
  • _new_do: called for object creation
  • _init_pre: called after object creation / before object initialization
  • _init_do: called fo object initialization
  • _init_post: called after object initialization

Usage:

A quick example which filters None and arguments which do not convert to float from the kwargs before the object is created:

from metaframe import MetaFrame


class A(MetaFrame.as_metaclass(object)):

    @classmethod
    def _new_do(cls, *args, **kwargs):

        nkwargs = dict()
        for key, val in kwargs.items():

            # Remove any argument with a value of None
            if val is None:
                continue

            try:
                val = float(val)
            except:
                continue

            nkwargs[key] = val

        # The only nuisance being the cumbersome call to _new_do
        # super doesn't work
        obj, args, kwargs = cls.__class__._new_do(cls, *args, **nkwargs)
        return obj, args, kwargs

    def __init__(self, **kwargs):
        for key, val in kwargs.items():
            print('key, val, type', key, val, type(val))


a = A(p1=72, p2=None, p3='hello', p4=None, p5='72.5')


# Now with a subclassed MetaB from MetaFrame
# Here super can be applied to find the higher in the hierarchy _new_do

class MetaB(MetaFrame):

    def _new_do(cls, *args, **kwargs):

        nkwargs = dict()
        for key, val in kwargs.items():

            # Remove any argument with a value of None
            if val is None:
                continue

            try:
                val = float(val)
            except:
                continue

            nkwargs[key] = val

        # super can be called directly
        obj, args, kwargs = super(MetaB, cls)._new_do(*args, **nkwargs)
        return obj, args, kwargs


class B(MetaB.as_metaclass()):
    def __init__(self, **kwargs):
        for key, val in kwargs.items():
            print('key, val, type', key, val, type(val))


b = B(p1=27, p2=None, p3='olleh', p4=None, p5='5.27')