Interesting Python Problems

Metaclass Method Inheritance.py

class BaseMeta(type):
    def __init__(cls, *args):
        cls.f = cls.f  # comment this line and run again!
        pass

    def f(cls):
        print(cls)

class Derived(metaclass=BaseMeta):
    pass


Derived().f()  # <class '__main__.Derived'>  # gives AttributeError for the other case!
# Derived.f()  # works for both cases

# https://chatgpt.com/share/67e17ede-f60c-800f-b4b6-60a086c6f9bf
# Without the statment `cls.f = cls.f`, the `Derived` class has the classmethod `f()` as an atttribute that is called using the
# `__getattribute__()` function; otherwise, it explicitly copies the method to the `Derived` class `__dict__`, so it available
# to its instances as well.

Numpy Array Comparisons.py

import numpy as np


a = np.array([1]), np.array([1])
assert np.array_equal(a, a)  # True

a = np.array([1, 1]), np.array([1])
assert np.array_equal(a, a)  # False!

# https://chatgpt.com/share/67e184e0-a2c4-800f-b8fe-566d4f603b6d
# `np.array_equal` forms a regular multi‐dimensional array when the subarrays are of the same size, allowing element‐wise comparison.
# With different sizes, it falls back to an object array where equality isn’t defined element‐wise, so the comparison becomes ambiguous.