TypeError: unsupported operand type(s) for |: 'type' and 'NoneType'

[wizard-28]

First Check

• I added a very descriptive title to this issue.
• I used the GitHub search to find a similar issue 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 __future__ import annotations

import typer

app = typer.Typer()


def main():
    app()


@app.command()
def test(flag: bool | None = typer.Option(None, "--flag", "-f", help="A flag")):
    print(flag)


if __name__ == "__main__":
    main()

Description

1. Run the example code
2. See the error: https://sourceb.in/SxyHEeLZxf

Operating System

Linux

Operating System Details

No response

Typer Version

0.4.0

Python Version

Python 3.8.10

Additional Context

No response

[ankitpv]

| was added in python 3.10, so not available in python 3.8.

def test(flag: bool | None = typer.Option(None, "--flag", "-f", help="A flag")):

Either you can use typing.Union or you can use string for type annotation, like:

def test(flag: "bool | None" = typer.Option(None, "--flag", "-f", help="A flag")):

[wizard-28]

| was added in python 3.10, so not available in python 3.8.

def test(flag: bool | None = typer.Option(None, "--flag", "-f", help="A flag")):

Either you can use typing.Union or you can use string for type annotation, like:

def test(flag: "bool | None" = typer.Option(None, "--flag", "-f", help="A flag")):

---

from __future__ import annotations

I've imported it from __future__ though.

Also, this sample code works in python 3.8, as I've imported annotations

from __future__ import annotations


def test(arg: int | str) -> None:
    print(arg)


test("string")
test(999)

[astrochun]

I'm getting the same issue with python 3.10:

RuntimeError: Type not yet supported: str | None

[jrmidkiff]

def create(self, 
        major: int|str, minor: int|str, patch: int|str, 
        prerelease: str, buildmetadata: str) -> 'Version': 
    pass

This causes TypeError: unsupported operand type(s) for |: 'type' and 'type' when I import this file into ipython, but not when I run this file from the command line or start a python session from the command line and import the file that way.
Python version 3.10.4

Turns out my iPython was pointing to wrong python version (3.7.10) - Needed to point it to 3.10

[parallaxe]

Would the typealias

String = str | None

be a possibility? It should match the semantics. The other types (Float, ...) could also be turned into aliases.

Though, Typealias seems to require python 3.10 (https://peps.python.org/pep-0613/).
Requiring 3.10 would be no issue for me as I'm already developing with it.

Is supporting < 3.10 a hard requirement for this module?

[johnthagen]

Duplicate of

• PEP 604 Support #348

[svlandeg]

Confirmed that either #676 or #548 will fix this issue!

[Asthestarsfalll]

Confirmed that either #676 or #548 will fix this issue!

Hi, When will those PRs merge?

[svlandeg]

Confirmed that either #676 or #548 will fix this issue!

Hi, When will those PRs merge?

When Tiangolo has a chance to review them - he's aware of them, but has many other things to do as well 😉

[alanwilter]

Oh, I need this fix please! I started refactoring to use typer but I now I'm on hold again.

[blackpearl1022]

def to_numpy(x: torch.Tensor | np.ndarray | dict | list) -> np.ndarray: TypeError: unsupported operand type(s) for |: 'torch._C._TensorMeta' and 'type'

How to do in this case ?