I use TypedDict with JSON so I get good hints in my IDE when dealing with complex JSON structures coming in from web APIs.

This includes hints of key names as well as methods to use with the types of data in the structure.


from typing import List, TypedDict


class SubDataTypedDict(TypedDict):
    id: int
    name: str
    widgets: List[int]


class DataTypedDict(TypedDict):
    id: int
    subDatas: List[SubDataTypedDict]
    foo: str


def data_doer(data: DataTypedDict):
    # using `data` in this function will give me completion on key names
    # as well as being aware of the types of the values and giving me 
    # completion on relevant methods.

On Wed, May 20, 2020 at 2:52 PM Luciano Ramalho <luciano@ramalho.org> wrote:
I hope you're all well.

I am writing the 2nd edition of Fluent Python and right now I am
trying to wrap up a section on typing.TypedDict, but I am having a
very hard time understanding its value in real world scenarios.

PEP 589 acknowledges that JSON is a "canonical use case" for dicts
used as records. But since the whole point of a TypedDict is to
provide syntax for type hints, I can't find a connection or any
benefit of using TypedDict with code that deals with JSON—where all
the action is at runtime.

All the examples in the PEP use literal dicts as values when assigning
to variables annotated with a TypedDict type.

If I annotate a function parameter with a TypedDict, and I am forced
to spell out all the keys to provide a dict argument that can be type
checked, I don't see why not use keyword arguments in the function
definition in the first place...

My testing shows that Mypy is unable to do much validation beyond
assignment from dict displays or dict constructor calls with keyword
arguments.

More realistic use cases that I was able to imagine often need casts
and don't perform any useful static checking.

My future readers and I could really use more insight about real world
use cases where TypedDict adds value.

Thanks for any help!

Cheers,

Luciano


--
Luciano Ramalho
|  Author of Fluent Python (O'Reilly, 2015)
|     http://shop.oreilly.com/product/0636920032519.do
|  Technical Principal at ThoughtWorks
|  Twitter: @ramalhoorg
_______________________________________________
Typing-sig mailing list -- typing-sig@python.org
To unsubscribe send an email to typing-sig-leave@python.org
https://mail.python.org/mailman3/lists/typing-sig.python.org/
Member address: dustin.wyatt@gmail.com