dynamically adding functions to a Python module

问题

Our framework requires wrapping certain functions in some ugly boilerplate code:

def prefix_myname_suffix(obj):
    def actual():
        print 'hello world'
    obj.register(actual)
    return obj

I figured this might be simplified with a decorator:

@register
def myname():
    print 'hello world'

However, that turned out to be rather tricky, mainly because the framework looks for a certain pattern of function names at module level.

I've tried the following within the decorator, to no avail:

current_module = __import__(__name__)
new_name = prefix + func.__name__ + suffix
# method A
current_module[new_name] = func
# method B
func.__name__ = new_name
current_module += func

Any help would be appreciated!

回答1:

use either

current_module.new_name = func

or

setattr(current_module, new_name, func)

回答2:

It seems the solution to your problem would be to make the decorated function act as the original function.

Try using the function mergeFunctionMetadata from Twisted, found here: twisted/python/util.py

It makes your decorated function act as the original, hopefully making the framework pick it up.

来源：https://stackoverflow.com/questions/1621350/dynamically-adding-functions-to-a-python-module