Question

I’ve been learning meta-programming in Ruby, and have found it quite useful. I’m sure I can do the same in Python.

E.g.: how would I rewrite this function given this function in a concise and general way using meta-programming?

def foo(bar=None, baz=None, qux=None, haz=None):
    txt = {}

    if bar:
        txt.update({'bar': bar})
    if baz:
        txt.update({'baz': baz})
    if qux:
        txt.update({'qux': qux})
    if haz:
        txt.update({'haz': haz})

    return txt

(this is obviously over-simplified, in practice one might perform different tasks based on value-set of individual keys)

Answer 1

You could use the **kwargs syntax:

def foo(**kwargs):
    txt = {}
    for k, v in kwargs.items():
        if v:
            txt[k] = v
    return txt

Or, if you’re short on ink:

def foo(**kwargs):
    return dict(((k,v) for k, v in kwargs.items() if v))

As suggested by Clement Bellot:

If you want to be sure those args are in your expected set, you could have an ALLOWED_PARAMS = ('bar','baz','qux','haz') constant and replace if v with if v and k in ALLOWED_PARAMS.

As suggested by DSM:

If you’re running Python 2.7+, you can use a dict comprehension.
This version also offers an alternate take on the “Allowed variables” problem:

def foo(bar=None, baz=None, qux=None, haz=None): 
    return {k:v for k,v in locals().items() if v}

---

Let’s say you actually wanted to call functions depending on those parameters being there, you could do:

KWARGS_TO_FUNCTIONS = {
    'bar':fn_bar,
    'baz':fn_baz, 
    'qux':fn_qux,
    'haz':fn_haz,
}

def foo(**kwargs):
    for k, v in kwargs.items():
        if v:
            try:
                KWARGS_TO_FUNCTIONS[k](v)
            except KeyError:
                pass # We didn't want that kwarg.

---

Now, all in all, I’m not totally sure that’s really ‘meta’.