Type-Checking Python Programs With Type Hints and mypy

dbader.org/python-tricks ► Avoid the bugs that Python type hinting will catch in the first place and see how to write clean and Pythonic code
Get quick intro to Python type hinting using the mypy type checker, in order to catch bugs and improve the quality of your existing Python code.
In this video I'll walk you through a simple example of what Python's new type hints allow you to do, then I'll show you how to install and use the mypy type checker as a static analysis tool from the command line. At the end you'll get an outlook of the future of type checking and the type hinting syntax in Python.
FREE COURSE - "5 Thoughts on Mastering Python" dbader.org/python-mastery
SUBSCRIBE TO THIS CHANNEL: dbader.org/youtube
* * *
► Python MUGS, T-SHIRTS & MORE: nerdlettering.com
FREE Python Tutorials & News:
» Python Tutorials: dbader.org
» Python News on Twitter: / @dbader_org
» Weekly Tips for Pythonistas: dbader.org/newsletter
» Subscribe to this channel: dbader.org/youtube

[@FRIENDSofCAP]

FYI:: mypy-lang is not longer supported according to the terminal response... And as of pyton v3.7.4 (which I am running) suggests installing 'mypy' (i.e., pip install mypy'

[@eduardoruizrios6140]

Maybe few years later, but now as indicated in the official documentation you can install mypy by running the following command in your command-line tool: python3 -m pip install -U mypy. In addition, is not mandatory to run the function for mypy to warning you about type errors. Actually you could just run mypy file_name.py and it will display on the command line all the lines where a type error was found!

[@RameenFallschirmjager]

From the chaos and the menace that rage out there, I take shelter in serenity and stability and reliability of programming!

[@hehehe289]

two things:
1, you should run pip install mypy now since the package names have been updated
2, you don't have to run the function in your code for it to be checked by mypy

[@rikless]

Is it possible to type hint objects ? For example, to force a specific class type on an argument.

[@MrNagios]

Thanks for explanation.

[@EmadMokhtar]

Is there a way to integrate MyPy to test suite?

[@lindhe]

Cool. Does type hints work with keyword arguments also?

[@OthmanAlikhan]

Thanks for the video =)

[@aoeu256]

IMO: doctests with extended syntax can be used to derive types, and they are a better form of documentation than types.
Which is more useful:
join(Sequence[Sequence]) -> Sequence
or
% join(["cold", "bath"]) -> "coldbath"
% intadd(!1, !2) -> !3
So it will look at the arguments in your test, and annotate the function but any list or str will generalize into a Sequence and ints and floats will turn to Number unless you write ! which means don't try to "generalize" the types. It can run at about the same time that the type system works on.

[@sinekonata]

So 4 years later, is type-hinting and type checking the confirmed future of python?

[@dhaval1489]

Does it help decrease processing time?, Does the interpreter check hints and not spend time on guessing type on run time? That would be huge incentive other I am good with dynamic style.

[@user-nf8zb4qp6j]

AFAIK, no. This is not really affect your program in a runtime. It's just helps to check that you are using your function with a right types. You still can call function with arguments of any types.

[@tcarney57]

Two questions and a comment: How would mypy react if you left out the annotations on the arguments but included the one on the return value i.e., *-> int:* ? Can mypy read annotations in a same-line comment instead of embedded in the code?
My comment is that embedded hints add some visual noise for the human reader. *add_them(a, b):* is quite a bit easier for the eye to follow than *add_them(a: int, b: int) ->:* . A compromise might be:
*add_them(a, b) # a: int, b: int, -> int*
In the good-old qbasic days, you used a single character ($, %, etc.) appended to the variable name to force a type: ex. *name$* or *age%* . This was not as visually disconcerting as the PEP484 annotations. I guess as long as one reads annotated python code in a color-coded editor, the visual noise will be reduced somewhat since the annotations will be immediately apparent.

[@vincenzofiorentini8002]

for me, this works in 3.5.4 by having an 'import mypy' in the code. with it, the type check works, without it has no effect. 'mypy-lang' seems seriously deprecated.

[@aharongina5226]

we have inspect functionality in pycharm

[@vijayabhaskar-j]

What if I need the add_this function to handle both int and float?

[@Berl0ga]

Vijayabhaskar J Union[int, float] or Number (from typing)

[@krishna_o15]

does this needs to be done for every overridden method also.

[@1mayanksuri]

yes we need to do

[@soldadopreciso]

thanks pal.

[@ORagnar]

I just realized, you sound like the Yankee slugger Alex Rodriguez.

[@jonathanmoore5619]

Python is starting to eat its own tail... Programming is about a means to an end... Not an end in itself.

[@sgt391]

it is an end for people developing the language

[@RameenFallschirmjager]

So vanilla python is not that much great after all!