Question

I would like to create a python directory whose values are to be evaluated separately.

So, for example, in the following non-working example I define

a = {'key1': 'value1', 'key2': 42, 'key3': foo(20)}

for which e.g.

def foo(max):
    """Returns random float between 0 and max."""
    return max*random.random()

and after processing of the dict I want to have e.g.

a_processes = {'key1': 'value1', 'key2': 42, 'key3': 12.238746374}

The example does not work since the value of key key3 is immediately evaluated, and foo(20) is no callable. The way it could work would be to use something like

a = {'key1': 'value1', 'key2': 42, 'key3': foo}

but here foo will miss its arguments. One way to handle this would be to define the dict in the following way

a = {'key1': 'value1', 'key2': 42, 'key3': [foo, 20]}

with the following processing scheme

a_processed = dict([k,process(v)] for k,v in a.items())

in which process is a meaningful function checking if its argument is a list, or if the first element is a callable which gets called with the remaining arguments.

My question is about a better way/idea to implement my basic idea?

Answer 1

I often use an argument-free lambda here, just because I like the syntax:

>>> def foo(x):
...     return x*100
... 
>>> a = {'key1': 'value1', 'key2': 42, 'key3': lambda: foo(20)}
>>> a
{'key3': <function <lambda> at 0x96ae80c>, 'key2': 42, 'key1': 'value1'}
>>> {k: v() if callable(v) else v for k,v in a.iteritems()}
{'key3': 2000, 'key2': 42, 'key1': 'value1'}