Support Unpack aka PEP 692 annotations

[multimeric]

First Check

• I added a very descriptive title here.
• I used the GitHub search to find a similar question and didn't find it.
• I searched the Typer documentation, with the integrated search.
• I already searched in Google "How to X in Typer" and didn't find any information.
• I already read and followed all the tutorial in the docs and didn't find an answer.
• I already checked if it is not related to Typer but to Click.

Commit to Help

• I commit to help with one of those options 👆

Example Code

from typing import Unpack, TypedDict
import typer

class Args(TypedDict):
    a: int
    b: str

def main(**kwargs: Unpack[Args]):
    pass

typer.run(main)

Traceback (most recent call last):

  File "/private/var/folders/q5/fvpyppm924525yl9m3xpgjm400031g/T/tmp.dvMKlgyv/test.py", line 11, in <module>
    typer.run(main)

RuntimeError: Type not yet supported: *<class '__main__.Args'>

Description

Above is a fairly simple example of how Unpack might be used with typer. There are many reasons why a user might want to do this, as outlined in the PEP. My immediate use case for this is to re-use an argument definition I use elsewhere, so I don't have to duplicate the function's type signature and risk the duplicated copies getting out of sync.

Typer would be able to support this by introspecting the TypedDict using __optional_keys__ etc.

Operating System

Linux, macOS

Operating System Details

Monterey (12.6)

Typer Version

0.9.0

Python Version

3.11.5

Additional Context

No response

[elupus]

This would solve a lot of argument re-use requirements in #153

[nronnei]

Also running into this, and would love to have it. Would be really nice to be able to define a base set of cli options to include on all methods. There's a lot going on in #153 that would be greatly simplified with PEP 692 support.

[aantn]

Strong yes from us! We're using typer for github.com/robusta-dev/holmesgpt/ and github.com/robusta-dev/krr/ and this would solve some annoying problems.

[aantn]

Giving this more thought, I believe that TypedDict doesn't allow something like this:

class Args(TypedDict):
    a: int = typer.Option(...)

However Annotated should work fine, which is an acceptable compromise I think

[aantn]

Yep. We don't use it in our projects today despite the official recommendation because it's lesser known syntax and we don't want to intimidate first time contributors who aren't python experts in favor of something that has no real benefits for us. But being able to reuse arguments between commands would be reason enough to switch. So still +1 from me

[nronnei]

I started an attempt at adding support for TypedDict & Unpack and got a fair bit of the way along, but I'm not good enough in Python to feel confident about my choices. Either of you open to looking at it?

[multimeric]

Share the PR and I might be able to have a look.

[aantn]

@nronnei any updates? Very interested in trying it

[nronnei]

Sorry for the very long delay, @aantn. Here's my very WIP attempt: master...nronnei:typer:nronnei/add-kwargs-typeddict-support

I haven't been able to get back to this since I'd started.