将多个函数包含在同一个语句中的pythonic方法是什么？

Question

I am using the Python library, Fabric , to do some remote server maintenance. Fabric automatically outputs all of the responses to remote and local commands unless you wrap the command in a couple with statements. Like so, on a local machine,

with settings(warn_only='true'):
    with hide('running', 'stdout', 'stderr', 'warnings'):
        output = local("uname -a", True)

or like this on a remote machine:

with settings(warn_only='true'):
    with hide('running', 'stdout', 'stderr', 'warnings'):
        output = run("uname -a")

I am writing a long and complex task and find myself repeating those two with statements over and over again. I want to write a function called _mute() to prevent that repetition. It would let me do something like this:

def _mute(fabric_cmd, args):
    with settings(warn_only='true'):
        with hide('running', 'stdout', 'stderr', 'warnings'):
            output = fabric_cmd(args)
    return output

def some_remote_task():
    # Run a remote task silently
    _mute(remote, 'uname -a')

def some_local_task():
    # Run a local task silently
    _mute(local, 'uname -a', True)

I've looked into some solutions and know that "eval" could do this for me. But every page I read about eval suggests that it's almost always a bad idea because of security issues. I looked into partials, but I couldn't figure out how to make an argument in my _mute function callable. I'm guessing there's a higher level Python concept I'm missing here. What's the pythonic way to go about doing this? Thanks for any direction you might be able to provide.

Answer 1

The better solution would be for you to build your own context manager; by far the easiest way would be to use the contextlib.contextmanager decorator :

from contextlib import contextmanager

@contextmanager
def _mute():
    with settings(warn_only='true'):
        with hide('running', 'stdout', 'stderr', 'warnings'):
            yield

Then use _mute as a context manager:

def some_remote_task():
    # Run a remote task silently
    with _mute():
        output = remote("uname -a")

This is a lot more compact and readable than having to retype the two larger context manager lines and has the added advantage that now you can run multiple commands in that same context.

As for your question; you can easily apply arbitrary arguments to a given function using the *args syntax:

def _mute(fabric_cmd, *args):
    with settings(warn_only='true'):
        with hide('running', 'stdout', 'stderr', 'warnings'):
            return fabric_cmd(*args)

def some_remote_task():
    # Run a remote task silently
    output = _mute(remote, 'uname -a')

See *args and **kwargs? for more information on the *args arbitrary argument lists tricks.