Simpler and simpler data class definitions: named tuple

Ashamed to say, I have been using python for a long time, and this is the first time I found such a good thing as namedtuple. Code first:

from collections import namedtuple
Point = namedtuple('Point', ['x', 'y'])
pt1 = Point(1.0, 5.0)
pt2 = Point(2.5, 1.5)

from math import sqrt
line_length = sqrt((pt1.x-pt2.x)**2 + (pt1.y-pt2.y)**2)

Isn't it fun? Point is like a class that has defined x and y properties. It can directly create instances, and all instances can access its properties in the form of .x and .y, which is much easier than defining classes directly:

class Point:
    def __init__(self, x: float, y: float):
        self.x = x
        self.y = y


pt1 = Point(1.0, 5.0)
pt2 = Point(2.5, 1.5)

from math import sqrt

line_length = sqrt((pt1.x - pt2.x) ** 2 + (pt1.y - pt2.y) ** 2)

Obviously, it's also more readable than using a raw tuple:

pt1 = (1.0, 5.0)
pt2 = (2.5, 1.5)

from math import sqrt
# use index referencing
line_length = sqrt((pt1[0]-pt2[0])**2 + (pt1[1]-pt2[1])**2)
 # use tuple unpacking
x1, y1 = pt1

There are only two members, and if there are more, the various numbers will confuse people.

So namedtuple is a shortcut for creating data classes based on tuple, allowing us to define data structures quickly and easily. Of course, it also has an obvious disadvantage: since the bottom layer is implemented with tuple, it is impossible to change (immutable), so pt1.x = 7 such as 061e150b94beb3 will obviously report an error.

If you want a mutable shortcut data class, you can use pyrecord:

from pyrecord import Record

Point = Record.create_type('Point', 'x', 'y')
pt1 = Point(1.0, 5.0)
pt2 = Point(2.5, 1.5)

from math import sqrt

line_length = sqrt((pt1.x - pt2.x) ** 2 + (pt1.y - pt2.y) ** 2)

pt1.x = 3.6
new_length = sqrt((pt1.x - pt2.x) ** 2 + (pt1.y - pt2.y) ** 2)