I'd like to define some contextual functions that can then be used later, as per the example below.
def add(x, y):
return x + y
def g(x):
u = my_func(x, 4 ** x)
return my_func(u, u // 2)
eval('g(1)', globals() | {'my_func': add})
This code throws NameError: name 'my_func' is not defined. Is there any way to achieve this (ideally without changing the functions add and g)?
When a function executes it looks up a name in the global namespace referenced by the function's __globals__ attribute, rather than the global namespace in which the call is made, so you would have to inject my_func into the function's __globals__ dict instead:
def add(x, y):
return x + y
def g(x):
u = my_func(x, 4 ** x)
return my_func(u, u // 2)
g.__globals__['my_func'] = add
eval('g(1)')