I'm trying to understand the following about dealing with functions and their arguments:
def print_my_arg(func, *args, **kwargs):
    func(*args, **kwargs)
    if 'my_arg' in kwargs: 
        print('  my_arg = {}'.format(kwargs['my_arg']))
def foo(call_no, my_arg='Default Value'):
    print('call_no = {}'.format(call_no) )
print_my_arg(foo, 0, my_arg='My Value 1')
print_my_arg(foo, 1, 'My Value 2')
print_my_arg(foo, 2)
Output:
call_no = 0
  my_arg = My Value 1
call_no = 1 # I'd like to see 'My Value 2' here
call_no = 2 # I'd like to see 'Default Value' here
Obviously people are free to invoke functions in either of the ways shown above, which makes me wonder: why my_arg doesn't go to kwargs anyway? Isn't there a uniform way to access parameters by name (and not by position), which doesn't depend on the way the function was invoked? 
Please note that:
- I'm not interested in - print_my_args(func, call_no, my_arg), because I'm talking about the case where I don't know the signature of- funcin advance and yet I want to know if a particular parameter exists (by name).
- Clearly that's related to decorators, but I've written the example in a simpler way (or I hope so). 
EDIT
Many thanks for the answers about inspect.signature. Using that, my new version of print_my_arg() is:
from inspect import signature
def print_my_arg ( func, *args, **kwargs ):
  func ( *args, **kwargs )
  sig = signature ( func )
  if 'my_arg' not in sig.parameters: return
  binding = sig.bind ( *args, **kwargs )
  binding.apply_defaults ()
  print ( "  my_arg = {}".format ( binding.arguments [ 'my_arg' ] ) )
 
    