Fair point!

@@Indently great post, though, thanks.

​@@Indently Thanks for every thing❤

​@@Indently please tutorial full all method

Or better yet, casefold to deal with some non-ascii characters. Possibly unicode normalization as well for stuff like combining characters. Text is hard.

G-strings can show a lot too.

​@@davidmurphy563 wtheckk😂

Since which version? And is it pronounced bang r? Does anyone remember back quotes being assign to repr?

@@DrDeuteron f-strings (or "formatted string literals") were added in *Python 3.6* , released end of 2016. The type converter !r to call repr() was there from the start, as proposed in PEP498 - along with !a for ascii(), and the rarely useful !s for str(). I have never thought about its pronounciation before, but I'm definitely going to use "banger" from now on. :D

@@nibblrrr7124 don't forget unicode() from 2x, but those backward quotes `foo` -> repr(foo) were weird. Though emacs knew about them,

Doesn’t type(self) work too?

Nice complement of multiplying strings! I might borrow that idea.

If you thought that was tough, imagine trying to do the same thing 15 to 20 years ago....

For me, I love overloading operators, until I want to see where something happens in the code.

Also it's weird to have "__or__" method that's not commutative. I would not expect `a | b == b | a` to return False and here it will. Edit: Fixed based on @mudi2000a comment.

certain functionality is fine for dunder methods, basically where they are self explainatory, __str__ is a simple expectation, __eq__ aswell an example where __add__ and other mathmatical operators is fine to overload is for example a Point class which just represents multiple numbers basically use dunder methods where its extremely clear what it does under the hood by just knowing what the class represents without looking at the actual implementation

I disagree with your first point (that it's not obvious that the "|" operator calls "__or__"), as that is just being familiar with Python. That ship has sailed with "__init__", hasn't it? (The fact that we have come to accept how messed-up and confusing constructor syntax in C++/Java is another topic...) However, I do agree that unless the meaning is obvious - like the interfaces from collections.abc, or behaving similarly to builtin datatypes ("+" for concatenation, like with str or list) - or well-documented and signposted, you're probably better off using regular methods and make the users write more explicit calls.

@@Raugturi You mean commutative.

One rule of thumb I try to follow: eval(repr(my_object)) == my_object So repr() should give the code to construct the same (or at least an equivalent) object. This is what e.g. dataclass or namedtuple do by default. Use dunder attributes for the class name, to avoid mistakes when renaming classes or copying code: def __repr__(self): return f"{__class__.__qualname__}(x={self.x!r}, ...)" If this one-liner gets too cluttered, put the class name in a helper variable cls and use ', '.join() or linebreak-separated strings within parentheses for the constructor args/attributes. Now that I think of it, this _could_ probably be automated into a class wrapper using introspection to get the constructor signature... I guess you could use name instead of qualname, b/c AFAIU it only makes a difference with nested classes, which IMHO should make anyone pause and question their choices, anyway. :^)

@@nibblrrr7124 Yeah, that's nice! The JSON string I usually emit is something like: {classname:{attributes and values}} Which does allow reconstruction of the object from the JSON and is fairly readable when debugging.

If you use it for debugging only, __repr__ is enough, it will be automatically called also instead of __str__ if __str__ does not exist.

What’s great about python is that you can test it in an interpreter about as fast as you can answer the question.

1) No, because == on the default collections (dict, list, set, ...) compares their contents by value. Use the *is* operator for checking for identity, i.e. whether they're literally referring to the same object in memory. 2) Yes, at least for normal user-defined classes (AFAIU classes can use slots instead of a class dict; also @property attributes might be an issue). Please don't, as it's hella confusing. But as Raymond Hettinger puts it: "Python is a language for consenting adults." (See also: Why are there no private methods?)

And you can return a different type regardless of the specification... 😄

@@sarimbinwaseem hmmm interesting

Every thing is a first class object in python

Don't think so (post python2 i presume?)

@@landsgevaer yes, I use 3.10 however I need compatibility to 3.7 and 3.8

Without checking: from __future__ import annotations and/or from typing import Self

It's the Røde NT USB

@@Indently merci

I just imagine an Easter egg in all of their code where if you add 'banana' to the end of an input, it will show an ASCII art banana for 5 seconds before moving on to the actual functionality of the code.

Luckily type hints don’t matter

begone javascript sympathizer. although python does not enforce it, you should not be using the == operator on objects of different types. if you want to compare objects and but are not sure the type, your problems are much deeper

Python document webpage

Google "list of all dunder methods in python" disappointed you?

Sounds German(ic).

No, but there is a mifflin

Pandas dataframes have a head method; not dunder though...

Fruit(name="apple & apple", grams=2000.0) the way he implemented it (I don't remember the weight).

Did you include the first line: from typing import Self Because I'm using Python 3.10, I had to include: from typing_extensions import Self

@@johnraz99 totally missed it. Thanks!

It's a type annotation. I was just playing around with the interpreter and it doesn't prevent me from assigning other types to a variable, so I'm not really sure what the point is other than to provide documentation for anyone reading the source code, but if you just call a constructor like that it's redundant for seemingly no reason. Maybe someone more knowledgeable about Python can correct us both, but for what it's worth, I wouldn't bother unless the declaration is separate from your first initialization of a variable.

@@anon_y_mousse oh got it! Yeah makes sense!

It doesn't prevent you of assigning anything else but it will show you a warning for it. Also it enables your IDE to include the appropriate methods that are associated with that type in the auto-complete list.

@@cookie_space I don't use an IDE. It's either vim or the python3 interpreter from the command line whenever I use it. The only time I've discovered that it actually errs is when used for parameter typing for functions, but I guess warnings will have to suffice for other uses. It would be nice if Python would use it to prevent assignment of unequal types, but in a language that treats variables as *really* variable, that's probably asking too much.

Yes, as someone else pointed, I only did half the job there xD

Thank you, I have updated both the thumbnail and description!

Do you mean the annotations? Like: Twopi : float = three: int + 0.1 : float + 41: int / pow(10: int, 3: int) ? Yes. And telling ppl dunder init returns None, I just can’t. Of dunder str returns str. Same for int, float, complex, etc….

@@DrDeuteron No. I just meant the IDE tooltips

@@user-bd1dh7hh1j oh. Yes, I turn those off all the time, esp when I'm writing a line that depends the prior line... ...it opens up and hides what I want to build off of and tells me a bunch of stuff I don't need to know.