How to initialize an object with a field a numpy a

问题:

This question is an exact duplicate of:

• How to make up lost reference to declare a field (numpy)? 2 answers

How to initialize an object with a field a numpy array, and the array is passed as an argument?

For example

class Foo
    def __init__()
        self.x = None
        self.m = None
        self.v = None

this has static methods

@staticmethod
def pop_x(x):
     # populate x with zeros
     x = zeros(n)

wouldn't work if I say self.pop(self.x) because set x= something it's just losing the referenced to x

def pop_x():
    # populate x with zeros
    self.x = zeros(n)

because I have a bunch of fields Object.x, Object.y, ... so I don't want to make a pop method for each of them.

@staticmethod
def update_q(q, i, val):
    """
    update i-th quantity
    """
    if q.ndim == 1:
        q[i] = val
    elif q.ndim == 2:
        q[i, :] = val

@staticmethod
def pop_q(q, n, m):
    """
    populate quantity with zeros
    """
    if q.ndim == 1:
        q = zeros(n)
    elif q.ndim == 2:
        q = zeros((n, m))

@staticmethod
def get_q(q, i):
    if q.ndim == 1:
        return q[i]
    elif q.ndim == 2:
        return q[i, :]

Here pop_q wouldn't work because setting q equals a numpy array lose the reference to q.

回答1:

This sounds like you may need a different design alltogether, e.g. separate the object that holds all those arrays from the object that does the calculations.

Maybe you could just put the arrays in a dict and pass the dict to your generic function, together with the name of the array, which would be the key in the dict?

Since the fields in a class are also in a dict, that approach can work with a class instance as well, using object.__dict__ and using string names instead of variables.