OOP：对象参数取决于我调用方法的位置

Question

This is not a purely technical issue, I am trying to find the cleanest and most efficient solution possible.

I am doing a tree exploration program in Python 3, and I use an object I need in each leaf of this tree. I want the values of the parameters of my object to be editable at different places in my program (in the different leaves), and the value to be updated in all the leaves, so I need to use the same object everywhere (pointers to the same object and not copies of the object). However, I would like one of the parameters to depend on the leaf in which the object is located.

I actually need many objects like that.

So I see several options:

• use two objects instead of one: an object A which contains an object B as parameter. So I make copies of object A for each new leaf, but the object B parameter always points to the same object (so my changes on it "propagate" to all leaves). So I can create my leaf-specific parameter in object A;
• store the leaf-specific variable separately from the object in each leaf, and send it as argument when I call a method of my object. This is less convenient because I use a lot of objects in each leaf, so I should have some kind of dictionary containing these variables whose keys are my objects. The problem is that some methods of my object are called from other objects, so I have to pass this dictionary through all called objects. It is less clean and the complexity seams greater (need to search in the dictionary every time I call a method);
• Same thing, but the dictionnary of leaf-specific parameters is stored in each object and the keys are the leaves. I have to give the current leaf pointer to each method I call.

I have never been faced to such a situation, so would like to have your opinion on these options or a better idea to write a code as clean and efficient as possible. Thanks for your help!

PS: I have about a thousand leaves and a hundred objects per leaf.

Currently working on option 1, still not sure if it's the best one.

Answer 1

Not super sure what you're asking. However, maybe class variables can solve your problem. Each Leaf has certain instance variables, that are specific/unique to each leaf instance. The Leaf class has a class variable which is shared among all leaf instances.

class Leaf:

    class_variable = 0

    def __init__(self):
        from random import randint
        Leaf.class_variable = randint(0, 100)
        self.instance_variable = randint(0, 100)


number_of_leaves = 5
leaves = [Leaf() for _ in range(number_of_leaves)]

for leaf in leaves:
    print(f"class variable: {leaf.class_variable}\ninstance variable: {leaf.instance_variable}")

Output:

class variable: 79
instance variable: 7
class variable: 79
instance variable: 57
class variable: 79
instance variable: 93
class variable: 79
instance variable: 2
class variable: 79
instance variable: 65
>>>