Pyright flaging mixed keys dictionary as an error

[Pilgrim1379]

The following is a copy paste from FastAPI - Query Parameters and String Validations

from typing import Annotated
from fastapi import FastAPI, Query

app = FastAPI()


@app.get("/items/")
async def read_items(q: Annotated[str | None, Query(max_length=50)] = None):
    results = {"items": [{"items_id": "Foo"}, {"item_id": "Bar"}]}
    if q:
        results.update({"q": q})
    return results

Pyright complains like below:

1. No overload for "update" match the provided arguments
2. Argument of type "dict[str, str]" cannot be assigned to parameter "m" of type "Iterable[tuple[str, list[dict[str, str]]]]" in function "update"
"Literal['q']" is not assignable to "tuple[str, list[dict[str, str]]]"

It seems pyright is overly strict, expecting that all values in results should match the structure of the items key (i.e., a list of dictionaries), when in fact you can update the dictionary with any new key-value pair (e.g., a string).

How can I configure pyright to not flag such line as an error?

[erictraut]

This is a legitimate type violation, so pyright is correct to generate an error here. You can suppress type errors with a # type: ignore or # pyright: ignore comments, but it's better to fix your code to eliminate the type violation. Otherwise you defeat the purpose of a type checker.

If your intent is for results to be a dict that accepts values of any type, you can declare its type as dict[str, object] by providing a type annotation.

results: dict[str, object] = ...

For more information about type declarations and type inference, refer to the pyright documentation.