관리-도구
편집 파일: decorators.py
"""This module implements decorators for implementing other decorators as well as some commonly used decorators. """ import sys PY2 = sys.version_info[0] == 2 if PY2: string_types = basestring, def exec_(_code_, _globs_=None, _locs_=None): """Execute code in a namespace.""" if _globs_ is None: frame = sys._getframe(1) _globs_ = frame.f_globals if _locs_ is None: _locs_ = frame.f_locals del frame elif _locs_ is None: _locs_ = _globs_ exec("""exec _code_ in _globs_, _locs_""") else: string_types = str, import builtins exec_ = getattr(builtins, "exec") del builtins from functools import partial from inspect import isclass from threading import Lock, RLock from .arguments import formatargspec try: from inspect import signature except ImportError: pass from .wrappers import (FunctionWrapper, BoundFunctionWrapper, ObjectProxy, CallableObjectProxy) # Adapter wrapper for the wrapped function which will overlay certain # properties from the adapter function onto the wrapped function so that # functions such as inspect.getargspec(), inspect.getfullargspec(), # inspect.signature() and inspect.getsource() return the correct results # one would expect. class _AdapterFunctionCode(CallableObjectProxy): def __init__(self, wrapped_code, adapter_code): super(_AdapterFunctionCode, self).__init__(wrapped_code) self._self_adapter_code = adapter_code @property def co_argcount(self): return self._self_adapter_code.co_argcount @property def co_code(self): return self._self_adapter_code.co_code @property def co_flags(self): return self._self_adapter_code.co_flags @property def co_kwonlyargcount(self): return self._self_adapter_code.co_kwonlyargcount @property def co_varnames(self): return self._self_adapter_code.co_varnames class _AdapterFunctionSurrogate(CallableObjectProxy): def __init__(self, wrapped, adapter): super(_AdapterFunctionSurrogate, self).__init__(wrapped) self._self_adapter = adapter @property def __code__(self): return _AdapterFunctionCode(self.__wrapped__.__code__, self._self_adapter.__code__) @property def __defaults__(self): return self._self_adapter.__defaults__ @property def __kwdefaults__(self): return self._self_adapter.__kwdefaults__ @property def __signature__(self): if 'signature' not in globals(): return self._self_adapter.__signature__ else: return signature(self._self_adapter) if PY2: func_code = __code__ func_defaults = __defaults__ class _BoundAdapterWrapper(BoundFunctionWrapper): @property def __func__(self): return _AdapterFunctionSurrogate(self.__wrapped__.__func__, self._self_parent._self_adapter) @property def __signature__(self): if 'signature' not in globals(): return self.__wrapped__.__signature__ else: return signature(self._self_parent._self_adapter) if PY2: im_func = __func__ class AdapterWrapper(FunctionWrapper): __bound_function_wrapper__ = _BoundAdapterWrapper def __init__(self, *args, **kwargs): adapter = kwargs.pop('adapter') super(AdapterWrapper, self).__init__(*args, **kwargs) self._self_surrogate = _AdapterFunctionSurrogate( self.__wrapped__, adapter) self._self_adapter = adapter @property def __code__(self): return self._self_surrogate.__code__ @property def __defaults__(self): return self._self_surrogate.__defaults__ @property def __kwdefaults__(self): return self._self_surrogate.__kwdefaults__ if PY2: func_code = __code__ func_defaults = __defaults__ @property def __signature__(self): return self._self_surrogate.__signature__ class AdapterFactory(object): def __call__(self, wrapped): raise NotImplementedError() class DelegatedAdapterFactory(AdapterFactory): def __init__(self, factory): super(DelegatedAdapterFactory, self).__init__() self.factory = factory def __call__(self, wrapped): return self.factory(wrapped) adapter_factory = DelegatedAdapterFactory # Decorator for creating other decorators. This decorator and the # wrappers which they use are designed to properly preserve any name # attributes, function signatures etc, in addition to the wrappers # themselves acting like a transparent proxy for the original wrapped # function so the wrapper is effectively indistinguishable from the # original wrapped function. def decorator(wrapper=None, enabled=None, adapter=None, proxy=FunctionWrapper): # The decorator should be supplied with a single positional argument # which is the wrapper function to be used to implement the # decorator. This may be preceded by a step whereby the keyword # arguments are supplied to customise the behaviour of the # decorator. The 'adapter' argument is used to optionally denote a # separate function which is notionally used by an adapter # decorator. In that case parts of the function '__code__' and # '__defaults__' attributes are used from the adapter function # rather than those of the wrapped function. This allows for the # argument specification from inspect.getfullargspec() and similar # functions to be overridden with a prototype for a different # function than what was wrapped. The 'enabled' argument provides a # way to enable/disable the use of the decorator. If the type of # 'enabled' is a boolean, then it is evaluated immediately and the # wrapper not even applied if it is False. If not a boolean, it will # be evaluated when the wrapper is called for an unbound wrapper, # and when binding occurs for a bound wrapper. When being evaluated, # if 'enabled' is callable it will be called to obtain the value to # be checked. If False, the wrapper will not be called and instead # the original wrapped function will be called directly instead. # The 'proxy' argument provides a way of passing a custom version of # the FunctionWrapper class used in decorating the function. if wrapper is not None: # Helper function for creating wrapper of the appropriate # time when we need it down below. def _build(wrapped, wrapper, enabled=None, adapter=None): if adapter: if isinstance(adapter, AdapterFactory): adapter = adapter(wrapped) if not callable(adapter): ns = {} # Check if the signature argument specification has # annotations. If it does then we need to remember # it but also drop it when attempting to manufacture # a standin adapter function. This is necessary else # it will try and look up any types referenced in # the annotations in the empty namespace we use, # which will fail. annotations = {} if not isinstance(adapter, string_types): if len(adapter) == 7: annotations = adapter[-1] adapter = adapter[:-1] adapter = formatargspec(*adapter) exec_('def adapter{}: pass'.format(adapter), ns, ns) adapter = ns['adapter'] # Override the annotations for the manufactured # adapter function so they match the original # adapter signature argument specification. if annotations: adapter.__annotations__ = annotations return AdapterWrapper(wrapped=wrapped, wrapper=wrapper, enabled=enabled, adapter=adapter) return proxy(wrapped=wrapped, wrapper=wrapper, enabled=enabled) # The wrapper has been provided so return the final decorator. # The decorator is itself one of our function wrappers so we # can determine when it is applied to functions, instance methods # or class methods. This allows us to bind the instance or class # method so the appropriate self or cls attribute is supplied # when it is finally called. def _wrapper(wrapped, instance, args, kwargs): # We first check for the case where the decorator was applied # to a class type. # # @decorator # class mydecoratorclass(object): # def __init__(self, arg=None): # self.arg = arg # def __call__(self, wrapped, instance, args, kwargs): # return wrapped(*args, **kwargs) # # @mydecoratorclass(arg=1) # def function(): # pass # # In this case an instance of the class is to be used as the # decorator wrapper function. If args was empty at this point, # then it means that there were optional keyword arguments # supplied to be used when creating an instance of the class # to be used as the wrapper function. if instance is None and isclass(wrapped) and not args: # We still need to be passed the target function to be # wrapped as yet, so we need to return a further function # to be able to capture it. def _capture(target_wrapped): # Now have the target function to be wrapped and need # to create an instance of the class which is to act # as the decorator wrapper function. Before we do that, # we need to first check that use of the decorator # hadn't been disabled by a simple boolean. If it was, # the target function to be wrapped is returned instead. _enabled = enabled if type(_enabled) is bool: if not _enabled: return target_wrapped _enabled = None # Now create an instance of the class which is to act # as the decorator wrapper function. Any arguments had # to be supplied as keyword only arguments so that is # all we pass when creating it. target_wrapper = wrapped(**kwargs) # Finally build the wrapper itself and return it. return _build(target_wrapped, target_wrapper, _enabled, adapter) return _capture # We should always have the target function to be wrapped at # this point as the first (and only) value in args. target_wrapped = args[0] # Need to now check that use of the decorator hadn't been # disabled by a simple boolean. If it was, then target # function to be wrapped is returned instead. _enabled = enabled if type(_enabled) is bool: if not _enabled: return target_wrapped _enabled = None # We now need to build the wrapper, but there are a couple of # different cases we need to consider. if instance is None: if isclass(wrapped): # In this case the decorator was applied to a class # type but optional keyword arguments were not supplied # for initialising an instance of the class to be used # as the decorator wrapper function. # # @decorator # class mydecoratorclass(object): # def __init__(self, arg=None): # self.arg = arg # def __call__(self, wrapped, instance, # args, kwargs): # return wrapped(*args, **kwargs) # # @mydecoratorclass # def function(): # pass # # We still need to create an instance of the class to # be used as the decorator wrapper function, but no # arguments are pass. target_wrapper = wrapped() else: # In this case the decorator was applied to a normal # function, or possibly a static method of a class. # # @decorator # def mydecoratorfuntion(wrapped, instance, # args, kwargs): # return wrapped(*args, **kwargs) # # @mydecoratorfunction # def function(): # pass # # That normal function becomes the decorator wrapper # function. target_wrapper = wrapper else: if isclass(instance): # In this case the decorator was applied to a class # method. # # class myclass(object): # @decorator # @classmethod # def decoratorclassmethod(cls, wrapped, # instance, args, kwargs): # return wrapped(*args, **kwargs) # # instance = myclass() # # @instance.decoratorclassmethod # def function(): # pass # # This one is a bit strange because binding was actually # performed on the wrapper created by our decorator # factory. We need to apply that binding to the decorator # wrapper function that the decorator factory # was applied to. target_wrapper = wrapper.__get__(None, instance) else: # In this case the decorator was applied to an instance # method. # # class myclass(object): # @decorator # def decoratorclassmethod(self, wrapped, # instance, args, kwargs): # return wrapped(*args, **kwargs) # # instance = myclass() # # @instance.decoratorclassmethod # def function(): # pass # # This one is a bit strange because binding was actually # performed on the wrapper created by our decorator # factory. We need to apply that binding to the decorator # wrapper function that the decorator factory # was applied to. target_wrapper = wrapper.__get__(instance, type(instance)) # Finally build the wrapper itself and return it. return _build(target_wrapped, target_wrapper, _enabled, adapter) # We first return our magic function wrapper here so we can # determine in what context the decorator factory was used. In # other words, it is itself a universal decorator. The decorator # function is used as the adapter so that linters see a signature # corresponding to the decorator and not the wrapper it is being # applied to. return _build(wrapper, _wrapper, adapter=decorator) else: # The wrapper still has not been provided, so we are just # collecting the optional keyword arguments. Return the # decorator again wrapped in a partial using the collected # arguments. return partial(decorator, enabled=enabled, adapter=adapter, proxy=proxy) # Decorator for implementing thread synchronization. It can be used as a # decorator, in which case the synchronization context is determined by # what type of function is wrapped, or it can also be used as a context # manager, where the user needs to supply the correct synchronization # context. It is also possible to supply an object which appears to be a # synchronization primitive of some sort, by virtue of having release() # and acquire() methods. In that case that will be used directly as the # synchronization primitive without creating a separate lock against the # derived or supplied context. def synchronized(wrapped): # Determine if being passed an object which is a synchronization # primitive. We can't check by type for Lock, RLock, Semaphore etc, # as the means of creating them isn't the type. Therefore use the # existence of acquire() and release() methods. This is more # extensible anyway as it allows custom synchronization mechanisms. if hasattr(wrapped, 'acquire') and hasattr(wrapped, 'release'): # We remember what the original lock is and then return a new # decorator which accesses and locks it. When returning the new # decorator we wrap it with an object proxy so we can override # the context manager methods in case it is being used to wrap # synchronized statements with a 'with' statement. lock = wrapped @decorator def _synchronized(wrapped, instance, args, kwargs): # Execute the wrapped function while the original supplied # lock is held. with lock: return wrapped(*args, **kwargs) class _PartialDecorator(CallableObjectProxy): def __enter__(self): lock.acquire() return lock def __exit__(self, *args): lock.release() return _PartialDecorator(wrapped=_synchronized) # Following only apply when the lock is being created automatically # based on the context of what was supplied. In this case we supply # a final decorator, but need to use FunctionWrapper directly as we # want to derive from it to add context manager methods in case it is # being used to wrap synchronized statements with a 'with' statement. def _synchronized_lock(context): # Attempt to retrieve the lock for the specific context. lock = vars(context).get('_synchronized_lock', None) if lock is None: # There is no existing lock defined for the context we # are dealing with so we need to create one. This needs # to be done in a way to guarantee there is only one # created, even if multiple threads try and create it at # the same time. We can't always use the setdefault() # method on the __dict__ for the context. This is the # case where the context is a class, as __dict__ is # actually a dictproxy. What we therefore do is use a # meta lock on this wrapper itself, to control the # creation and assignment of the lock attribute against # the context. with synchronized._synchronized_meta_lock: # We need to check again for whether the lock we want # exists in case two threads were trying to create it # at the same time and were competing to create the # meta lock. lock = vars(context).get('_synchronized_lock', None) if lock is None: lock = RLock() setattr(context, '_synchronized_lock', lock) return lock def _synchronized_wrapper(wrapped, instance, args, kwargs): # Execute the wrapped function while the lock for the # desired context is held. If instance is None then the # wrapped function is used as the context. with _synchronized_lock(instance if instance is not None else wrapped): return wrapped(*args, **kwargs) class _FinalDecorator(FunctionWrapper): def __enter__(self): self._self_lock = _synchronized_lock(self.__wrapped__) self._self_lock.acquire() return self._self_lock def __exit__(self, *args): self._self_lock.release() return _FinalDecorator(wrapped=wrapped, wrapper=_synchronized_wrapper) synchronized._synchronized_meta_lock = Lock()