Copy's Replace

00:00 In the previous lesson, I showed you the changes to docstring and the new command-line interface to the random module. This lesson covers the new replace method in the copy module.

00:11 copy module has a new function called replace, which copies an object but modifies one or more of its values in the copy. For example, say you want to copy a named tuple except for one of its attributes.

00:24 Before, you’d copy it, then make the change. Now you can just use the copy.replace() function instead. Let’s go look at this. Start by importing copy, and now to make that named tuple.

00:54 There’s more than one way to create a named tuple. The mechanism I’ve used here is similar to a data class. Now I can create a person and it’ll have three attributes: name, place, and Python version.

01:13 Let’s create a copy of person, but upgrade the Python version. The replace() function takes the object being copied, and then the keyword arguments specify what gets changed.

01:29 In this case, it’s the attribute Python version with a new value of 3.13, and it was a copy, so the original hasn’t changed. The original Python version still contains 3.12.

01:43 The logic for handling this copying is available in many of the common objects in the standard library. Let’s play with a date,

01:59 and now if I copy.replace() it, I can zero out the time part.

02:13 This feature gets handled under the covers using a dunder method called __replace__. This means you can implement your own version to specify how it behaves. Consider a custom Container class that has a name and contains a dictionary.

02:34 Here I’m using the keyword args feature of __init__, which passes any keyword arguments as a dictionary.

02:51 And then I store the name and dictionary for later. The __replace__ method uses **kwargs to specify the name-value pairs of the things to be replaced.

03:13 Inside the method, I’m responsible for creating the new copy, so I start out by getting an instance of the class, which I’m going to later use for construction.

03:29 Then I create a new dictionary, comprising the existing name and items. The or on the end here, uses any name-value pairs in the kwargs dictionary of the function, and overrides the corresponding values in init__kwargs.

03:45 This gives me a new dictionary with the replaced values. I use this to construct an object returning the new instance based on all of those keyword arguments.

04:02 To top it all off, I’ll implement a __repr__ method showing the contents of the container.

04:29 Alright, let’s create an instance of Container,

04:43 and now if I copy and replace it, you see the results. There’s two print statements from the __init__ and __replace__, and then the actual result is printed in the REPL through the __repr__.

05:01 Notice how the capital of Norway now is capitalized because I changed it using replace(). pathlib is one of my favorite modules, it’s been a big improvement on the original os module.

05:15 Python 3.13 adds more features to it, which I’ll show you in the next lesson.