Add Model.copy()

[JamesPHoughton]

It would be good to have a deep copy operation that can give you a copy of the current model to mess with. This could take the place of reloading a model after making changes, so you could take the copy after some changes, make other changes, and then drop them.

This could actually be pretty difficult, not sure.

I think a good syntax would be similar to that in networkx in that the copy function is a member of the object that it is copying. However, I expect the deepcopy is really what will be most useful.

[enekomartinmartinez]

Thinking about including this in 3.8.0

Model.copy() should be used as well for select_submodel. This function should create a copy and modify it, instead of mutating the original model.

[enekomartinmartinez]

I have been doing some test with the copy and I think it will not be so easy to implement as the calls inside the components are kept to the original data, e.g.:

 a -> b

c = copy(a)
d = copy (b)

we would like to c -> d, but we will have c -> b, running:

from copy import copy, deepcopy

def a():
    return b()

def b():
    return 2

c = copy(a)
d = deepcopy(a)
print(f"a = {a()}")
print(f"b = {b()}")
print(f"c = {c()}")
print(f"d = {d()}")
print("-"*5)

b = lambda: 5
print(f"a = {a()}")
print(f"b = {b()}")
print(f"c = {c()}")
print(f"d = {d()}")

returns

a = 2
b = 2
c = 2
d = 2
-----
a = 5
b = 5
c = 5
d = 5

The only option I see is that PySD tracks the changes done to the components, and the copy function loads a model again and applies the same changes.