Talk: Elizaveta Shashkova - The Hidden Power of the Python Runtime

[@bhupeshimself]

This was a really great talk. Thanks to everyone who helped in bringing this up ✨👍

[@elizavetashashkova893]

Thank you very much! :)

[@sitronco]

Thanks for your talk. It was a really interesting topic :)

[@elizavetashashkova893]

Thank you for your kind words! :)

[@NestiGX]

Great presentation :DD

[@elizavetashashkova893]

Thank you! :)

[@NestiGX]

I have a question regarding static type checkers. Tools like mypy, pyright, etc. are probably using functionalities mentioned by you in this presentation. I'm wondering how those tools could work, for example: I annotated my function like this "def fun(var: str) -> List[Dict[str, int]]". How can I check whether my function is returning that type or not? A simple type() call would only return information that variable is a list type. Also, we could probably gather full information about that variable only in run time, but yet those tools can analyze code without running the code.

[@elizavetashashkova893]

Hi Mateusz! As far as I know, these tools use only static analysis. So they don't run the code, they just parse the source code, build abstract syntax tree and control flow graph and try to derive types basing on this information. And, as you absolutely correctly said, in Python you usually can't predict 100% of information about types without code execution (because of the language nature). That's why I love Runtime Tools so much - you always know all the information about variables and program state :)

[@NestiGX]

Thank you a lot for explenation. ;)) Congratz on great presentation. I'm waiting for the next one :D